Laboratory Techniques in Biochemistry and Molecular Biology Vol 04

Laboratory techniques in biochemistry and molecular biology 4 LABORATORY TECHNIQUES IN BIOCHEMISTRY AND MOLECULAR BIO...

Author: Edited by T.S. Work and E. Work

424 downloads 2198 Views 20MB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
Report copyright / DMCA form

DOWNLOAD PDF

Laboratory techniques in biochemistry and molecular biology 4

LABORATORY TECHNIQUES IN BIOCHEMISTRY AND MOLECULAR BIOLOGY Volume 4 Edited by

T. S. WORK - N.I.M.R., Mill Hill, London E . WORK - Imperial College, London

Advisory board

G. POPJAK - U.C.L.A. S. BERGSTROM - Stockholm K. BLOCH Haruard University P. SIEKEVITZ - Rockefeller University E. SMITH U.C.L.A. E. C. SLATER - Amsterdam ~

~

NORTH-HOLLAND PUBLISHING COMPANY AMSTERDAM AMERICAN ELSEVIER PUBLISHING CO., INC. - NEW YORK ~

OXFORD

Part I

Part 11

A. N. Glazer, R. J. DeLange and D. S. Sigman CHEMICAL MODIFICATION OF PROTEINS Hannah Gould and H. R. Matthews SEPARATION METHODS FOR NUCLEIC ACIDS AND OLIGONUCLEOTIDES

1976 NORTH-HOLLAND PUBLISHING COMPANY - AMSTERDAM O X F O R D AMERICAN ELSEVIER PUBLISHING CO., INC. - NEW YORK

0 1976 North-Holland Publishing

Company

All rights reserved. N o part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, without the prior permission of the copyright owner.

series: 0 7204 4200 1 4: 0 7204 4215 X ISBN American Elsevier: 0 444 10985 4 Library of Congress Catalog Card Number: 68-54514 ISBN North-Holland

-

- volume

Published by: NORTH-HOLLAND PUBLISHING COMPANY - AMSTERDAM

Sole distributors for the U.S.A. and Canada: AMERICAN ELSEVIER PUBLISHING COMPANY, INC.

52 VANDERBILT AVENUE,

NEW YORK, N.Y.

Printed in The Netherlands

10017

. OXFORD

Editors’ preface

Progress in research depends upon development of technique. No matter how important the cerebral element may be in the planning of experiments, a tentative hypothesis cannot be converted into an accepted fact unless there is adequate consciousness of the scope and limitation of existing techniques; moreover, the results may be meaningless or even positively misleading if the technical ‘know how’ is inadequate. During the past ten or fifteen years, biochemical methods have become specialized and sophisticated to such a degree that it is now difficult for the beginner, whether undergraduate, graduate or specialist in another field, to grasp all the minor but important details which divide the successful from the unsuccessful experiment. In order to cope with this problem, we have initiated a new series of Laboratory Manuals on technique. Each manual is written by an expert and is designed as a laboratory handbook to be used at the bench. It is hoped that use of these manuals will substantially reduce or perhaps even remove that period of frustration which so often precedes the successful transplant of a specialized technique into a new environment. In furtherance of this aim, we have asked authors to place special emphasis on application rather than on theory; nevertheless, each manual carries sufficient history and theory to give perspective. The publication of library volumes followed by pocket paperbacks is an innovation in scientific publishing which should assist in bringing these manuals into the laboratory as well as into the library. In underV

VI

EDITORS’ PREFACE

taking the editing of such a diverse series, we have become painfully conscious of our own ignorance but have been encouraged by our board of advisers to whom we owe many valuable suggestions and, of course, by our authors who have co-operated so willingly and have so patiently tolerated our editoral intervention. T. S. & E. Work Editors

Contents of part I and I1

PART I

.

CHEMICAL MODIFICATION OF PROTEINSA . N . Glazer. R . J . DeLange and D . S. Sigman . . . . . . . . . . . . . . . . . . . . . . . . . . . . Contents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . List of abbreviations . . . . . . . . . . . . . . . . . . . . . . . . . . Chapter 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . Chapter 2. Chemical characterization of proteins and their derivatives . . . . Chapter 3. Modification of protein side-chains: group-specific reagents . . . . Chapter 4 . Site-specific modification of native proteins with group-specific reagents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Chapter 5 . Affinity labels . . . . . . . . . . . . . . . . . . . . . . . . Chapter 6. Photoaffinity labels . . . . . . . . . . . . . . . . . . . . . Appendix I . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Subject index . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1

3 9 10 13 68 121 135 167 180 189 201

PART I1 SEPARATION METHODS FOR NUCLElC ACIDS AND OLIGONUCLEOTIDES. Hannah Gould and H . R . Matthews . . . . . . . . . . . . . . . . . Contents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . List of abbreviations . . . . . . . . . . . . . . . . . . . . . . . . . . Oligonucleotide nomenclature . . . . . . . . . . . . . . . . . . . . . . Chapter 1 . Choice of method . . . . . . . . . . . . . . . . . . . . . . Chapter 2. Nucleotide composition of nucleic acids . . . . . . . . . . . . Chapter 3. Paper chromatography. paper electrophoresis and thin layer chromatography . . . . . . . . . . . . . . . . . . . . . . . . . . vii

207 209 215 217 219 225 241

VIIl

CONTENTS OF PARTS 1 A N D I1

Chapter 4 . Column chromatography . . . . . . . . . . . . . . . . . . Chapter 5 . Applications of column chromatography . . . . . . . . . . . Chapter 6. Desalting and concentrating oligonucleotide solutions . . . . . Chapter 7. Essentials of gel electrophoresis . . . . . . . . . . . . . . . Chapter 8. The practice of analytical gel electrophoresis of nucleic acids . . Chapter 9. Practice of preparative gel electrophoresis of nucleic acids . . . Chapter 10. Applications of gel electrophoresis . . . . . . . . . . . . . Chapter I1. Other techniques . . . . . . . . . . . . . . . . . . . . . . Appendix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Subject index . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

.

253

. 278

. 295 . 302 . 359

.

408

. 419 454 472 475 487

CHEMICAL MODIFICATION OF PROTEINS Selected methods and analytical procedures

A. N. Glazer, R. J. DeLange and D. S. Sigman Department of Biological Chemistry U.C.L.A. School of Medicine, Los Angeles, California, U.S.A.

This Page Intentionally Left Blank

Contents

List of abbreviations . . . . . . . . . . . . . . . . . . .

9

Chapter 1. lntroduction . . . . . . . . . . . . . . . . .

10

Chapter 2 . Chemical characterization of proteins and their derivatives . . . . . . . . . . . . . . . . . .

13

2.1. Amino acid analysis with special reference to problem amino acids and common derivatives . . . . . . . . . . . . . . . . . . . . . . . . 2.1.1. Preparation of proteins for hydrolysis . . . . . . . . . . . . . . 2.1.2. Acid hydrolysis of peptides and proteins . . . . . . . . . . . . . 2.1.3. Analysis and calculations . . . . . . . . . . . . . . . . . . . . 2.2. Serine. threonine, tyrosine and derivatives . . . . . . . . . . . . . . 2.2.1. Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2.2. Halogenated and nitrated derivatives of tyrosine . . . . . . . . . . 2.3. Valine and isoleucine . . . . . . . . . . . . . . . . . . . . . . . 2.4. Aspartic acid, glutamic acid, asparagine and glutamine . . . . . . . . . 2.5. Cysteine, cystine, methionine and derivatives . . . . . . . . . . . . . 2.5.1. Performic acid oxidation . . . . . . . . . . . . . . . . . . . . 2.5.2. Cysteic acid . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.5.3. S-Carboxymethylcysteine and S-carboxyethylcysteine . . . . . . . . 2.5.4. S-Aminoethylcysteine . . . . . . . . . . . . . . . . . . . . . 2.5.5. S-Methylcysteine . . . . . . . . . . . . . . . . . . . . . . . 2.5.6. Thialaminine . . . . . . . . . . . . . . . . . . . . . . . . . 2.5.7. S-Succinylcysteine (S-1,2-dicarboxyethyl-L-cysteine) . . . . . . . . 3

13 14 15 16 18

18 18 20 20 21 23 23 24 24 25 25 25

4

CHEMICAL MODIFICATIONS OF PROTEINS

2.5.8. Methionine sulfone . . . . . . . . . . . . . . . . . . . . . . 2.5.9. Methionine sulfoxides . . . . . . . . . . . . . . . . . . . . . 2.5.10. Methionine carboxymethylsulfonium salt . . . . . . . . . . . . 2.5.1 1. Homoserine and homoserine lactone . . . . . . . . . . . . . . 2.6. Tryptophan and derivatives . . . . . . . . . . . . . . . . . . . . . 2.7. Lysine derivatives . . . . . . . . . . . . . . . . . . . . . . . . . 2.7.1. E-N-Carboxymethyllysines . . . . . . . . . . . . . . . . . . . 2.7.2. Homoarginine . . . . . . . . . . . . . . . . . . . . . . . . 2.7.3. Deamination products of lysine . . . . . . . . . . . . . . . . . 2.7.4. Homocitrulline . . . . . . . . . . . . . . . . . . . . . . . . 2.8 Histidine and derivatives . . . . . . . . . . . . . . . . . . . . . . 2.8.1. Carboxymethylhistidines . . . . . . . . . . . . . . . . . . . . 2.8.2. Other derivatives of histidine . . . . . . . . . . . . . . . . . . 2.9. Arginine and derivatives . . . . . . . . . . . . . . . . . . . . . . 2.10. Other amino acids . . . . . . . . . . . . . . . . . . . . . . . . 2.1 1. Complete enzymic hydrolysis of proteins and peptides . . . . . . . . . 2.1 1.1. Procedure 1 : Pepsin, subtilisin, aminopeptidase M, prolidase . . . . 2.1 1.2. Procedure 2: Use of insoluble enzymes. . . . . . . . . . . . . . 2.1 1.3. Procedure 3: Papain, leucine aminopeptidase, prolidase . . . . . . 2.11.4. Analysis of enzymic hydrolysates . . . . . . . . . . . . . . . . 2.12. Naturally-occurring amino acid derivatives in proteins . . . . . . . . . . 2.12.1. Lysine derivatives . . . . . . . . . . . . . . . . . . . . . . . 2.12. I .1 . E-N-Acetyllysine . . . . . . . . . . . . . . . . . . . . . 2.12.1.2. e-N-Methyllysines . . . . . . . . . . . . . . . . . . . . . 2.12.1.3. Hydroxylysine . . . . . . . . . . . . . . . . . . . . . . 2.12.1.4. Trimethylhydroxylysine and its phosphate . . . . . . . . . . 2.12.1.5. Desmosine, isodesmosine and related derivatives . . . . . . . 2.12.1.6. Other lysine derivatives . . . . . . . . . . . . . . . . . . 2.12.2. Histidine derivatives . . . . . . . . . . . . . . . . . . . . . 2.12.3. Methylarginines . . . . . . . . . . . . . . . . . . . . . . . 2.12.4. 0-Phosphoserine and 0-phosphothreonine . . . . . . . . . . . . 2.12.5. Derivatives of tyrosine . . . . . . . . . . . . . . . . . . . . 2.12.6. Hydroxyprolines . . . . . . . . . . . . . . . . . . . . . . . 2.12.7. Other derivatives of amino acids . . . . . . . . . . . . . . . . 2.13. Analysis of carbohydrates in glycoproteins . . . . . . . . . . . . . . 2.13.1. Identification of a protein as a glycoprotein . . . . . . . . . . . 2.13.2. Release of carbohydrates from glycoproteins . . . . . . . . . . . 2.13.3. Analysis of carbohydrates in glycoproteins . . . . . . . . . . . . 2.14. End-group methods . . . . . . . . . . . . . . . . . . . . . . . 2.15. Sequential degradation and determination of amino-terminal residues (Edman degradation) . . . . . . . . . . . . . . . . . . . . . . . 2.16. Determination of COOH-terminal sequences . . . . . . . . . . . . .

26 27 29 29 30 34 34 34 35 35 36 36 37 37 38 38 40 41 42 43 43

44 44 45 47 48 48 49 49 50

51 52 52 53 53 54 55

57 59 60 64

CONTENTS

Chapter 3 . Modification of protein side-chains : group-specific reagents . . . . . . . . . . . . . . . . . . . 3.1. Modifications of amino groups . . . . . . . . . . . . . . . . . . . 3.1.1. Modifications of amino groups leading to retention of positive charge . 3.1.1.1. Amidination . . . . . . . . . . . . . . . . . . . . . . . 3.1.1.2. Guanidination . . . . . . . . . . . . . . . . . . . . . . . 3.1.1.3. Reductive methylation . . . . . . . . . . . . . . . . . . . 3.1.2. Modifications of amino groups resulting in conversion to a neutral derivative . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1.2. I . Carbamylation . . . . . . . . . . . . . . . . . . . . . . 3.1.2.2. Trifluoroacetylation . . . . . . . . . . . . . . . . . . . . 3.1.2.3. Trinitrobenzoylation . . . . . . . . . . . . . . . . . . . . 3.1.3. Conversion of amino groups to acidic derivatives . . . . . . . . . . 3.1.3.1. Succinylation . . . . . . . . . . . . . . . . . . . . . . . 3.1.3.2. Reversible modification of amino groups with maleic anhydride and similar reagents . . . . . . . . . . . . . . . . . . . . . . 3.2. Modifications of arginine . . . . . . . . . . . . . . . . . . . . . . 3.2.1. Butanedione reaction with arginine . . . . . . . . . . . . . . . . 3.2.2. Phenylglyoxal reaction with arginine . . . . . . . . . . . . . . . 3.2.3. Nitrornalondialdehyde reaction with arginine . . . . . . . . . . . 3.3. Modifications of carboxyl groups . . . . . . . . . . . . . . . . . . 3.4. Modifications of histidine . . . . . . . . . . . . . . . . . . . . . 3.5. Modifications of methionine . . . . . . . . . . . . . . . . . . . . 3.6. Modifications of tryptophan . . . . . . . . . . . . . . . . . . . . 3.6.1. Formylation . . . . . . . . . . . . . . . . . . . . . . . . . 3.6.2. Sulfenylation . . . . . . . . . . . . . . . . . . . . . . . . . 3.7. Modifications of tyrosine . . . . . . . . . . . . . . . . . . . . . . 3.7.1. Nitration . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.7.2. Iodination . . . . . . . . . . . . . . . . . . . . . . . . . . 3.8. Modifications of sulfhydryl and disulfide groups . . . . . . . . . . . 3.8.1. Modification by performic acid oxidation . . . . . . . . . . . . . 3.8.2. Reduction and carboxymethylation (or carboxamidomethylation) . . 3.8.3. Carboxyethylation of SH . . . . . . . . . . . . . . . . . . . . 3.8.4. Aminoethylation of SH . . . . . . . . . . . . . . . . . . . . . 3.8.5. Trimethylaminoethylation of cysteine . . . . . . . . . . . . . . . 3.8.6. Reduction and selective S-methylation . . . . . . . . . . . . . . 3.8.7. S-Sulfonation . . . . . . . . . . . . . . . . . . . . . . . . . 3.8.8. Conversion to the S-sulfenylsulfonate derivative . . . . . . . . . . 3.8.9. Addition of thiols to activated double bonds . . . . . . . . . . . . 3.8.9.1. Spectrophotometric titration with N-ethylmaleimide . . . . . . 3.8.9.2. Other applications of NEM and related compounds . . . . . . . 3.8.10. Titration of thiols with 5,5'-dithiobis( 2-nitrobenzoate) (DTNB;

5

68 69 69 69 70 71 73 73 74 76 78 78 79 84 84 86 87 88 89 90 91 92 93 95 95 99 101 102 103 104 105 106 107 108 109 110 110 111

6

CHEMICAL MODIFICATIONS OF PROTEINS

Ellman's reagent) . . . . . . . . . . . . . . . . . . . . . . . 3.8.1 1. Cyanylation of SH . . . . . . . . . . . . . . . . . . . . . . 3.8.12. Reaction of thiols with organomercurials . . . . . . . . . . . . . 3.8.12.1. Titration withp-mercuribenzoate (PMB) . . . . . . . . . . . 3.8.12.2. Reaction with mercurated nitrophenols . . . . . . . . . . . 3.8.13. Reaction of cysteinyl residues with azobenzene-2-sulfenyl bromide

113 114 116 116 117 119

Chapter 4 . Site-specijc modijcation of'nativeproteins with groupspecific reagents . . . . . . . . . . . . . . . . 121 4.1. Effect of complex formation . . . . . . . . . . . . . . . . . . . . 123 4.2. Effects of pK, perturbations . . . . . . . . . . . . . . . . . . . . 126 4.3. Widely employed site-specific reagents . . . . 4.3.1. Diisopropyl fluorophosphate . . . . . . 4.3.2. Pyridoxal-5-phosphate . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

130 130 131

Chapter 5 . Affinity labels . . . . . . . . . . . . . . . . . 135 5.1. Affinity labels versus group-specific reagents for site-specific modification . 5.2. Reactive groups for affinity labels . . . . . . . . . . . . . . . . . . 5.3. Haloketones . . . . . . . . . . . . . . . . . . . . . . . . . . 5.3.1. Haloketone derivatives of amino acids . . . . . . . . . . . . . . 5.3.1.1. Synthesis of haloketones by bromination . . . . . . . . . . . 5.3.1.2. Synthesis of haloketones by acidification of diazoketones with haloacids . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.3.2. Amino acid derivatives formed by haloketones . . . . . . . . . . . 5.3.2.1. Histidine . . . . . . . . . . . . . . . . . . . . . . . . . 5.3.2.2. Methionine . . . . . . . . . . . . . . . . . . . . . . . . 5.3.2.3. Serine . . . . . . . . . . . . . . . . . . . . . . . . . . 5.3.2.4. Aspartate and glutamate . . . . . . . . . . . . . . . . . . 5.3.2.5. Cysteine . . . . . . . . . . . . . . . . . . . . . . . . . 5.4. Amides and esters of haloacids . . . . . . . . . . . . . . . . . . . . 5.4.1. Synthesis of amides of haloacids . . . . . . . . . . . . . . . . . 5.4.2. Synthesis of esters of haloacids . . . . . . . . . . . . . . . . . 5.4.3. Analysis of acid hydrolysis products . . . . . . . . . . . . . . . 5.5. Epoxides . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.5.1. Synthesis . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.5.1.1.' Synthesis of glycidol phosphate from glycidol . . . . . . . . . 5.5.1.2. Synthesis of epoxides from alkenes . . . . . . . . . . . . . . 5.5.2. Analysis of products . . . . . . . . . . . . . . . . . . . . . . 5.6. Sulfonyl fluorides . . . . . . . . . . . . . . . . . . . . . . . . . 5.6.1. Synthesis of phenylmethylsulfonyl fluoride . . . . . . . . . . . . . 5.6.2. Synthesis of sulfanilyl fluorides . . . . . . . . . . . . . . . . . .

136 138 138 138 138

139 142 142 143 143 143 144 145 145 147 147 148 150 150 151 153 153

153 154

7

CONTENTS

5.6.3. Analysis of products . . . . . . . . . . . . . . . . . . . . . . 5.7. Diazo compounds-aryl diazonium salts . . . . . . . . . . . . . . . 5.7.1. Reactive properties of aryl diazonium salts . . . . . . . . . . . . 5.7.2. Synthesis of typical diazonium salts . . . . . . . . . . . . . . . 5.8. Diazo compounds-diazoketones and diazoacetates . . . . . . . . . . 5.8.1. Reactive properties . . . . . . . . . . . . . . . . . . . . . . 5.8.2. Synthesis of diazoacetyl derivatives . . . . . . . . . . . . . . . . 5.8.3. Analysis of products . . . . . . . . . . . . . . . . . . . . . .

156 157 157 160 162 162 164 165

Chapter 6 . Photoaffinity labels . . . . . . . . . . . . . . 167 6.1. Types of photoactivatable groups . . . . . . . . . . . . . . . . . . 6.1.1: Carbenes . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.1.2. Nitreoes . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.1.3. Benzophenones and acetophenones . . . . . . . . . . . . . . . . 6.2. Conditions of photolysis . . . . . . . . . . . . . . . . . . . . . . 6.2.1. Excitation wavelength . . . . . . . . . . . . . . . . . . . . . 6.2.2. Quantitative labelling . . . . . . . . . . . . . . . . . . . . . 6.2.3. Pseudo photoaffinity labelling . . . . . . . . . . . . . . . . . . 6.3. Synthesis of photoaffinity labels . . . . . . . . . . . . . . . . . . . 6.3.1. Synthesis of ethyl diazomalonyl derivatives . . . . . . . . . . . . . 6.3.2. Synthesis of aryl azides . . . . . . . . . . . . . . . . . . . . .

Appendix I .

. . . . . . . . . . . . . . . . . . . . . .

High voltage paper electrophoresis . . . . . . . . . . . . . . . Paper chromatography . . . . . . . . . . . . . . . . . . . . Paper chromatography or electrophoresis on a preparative scale . . Peptide mapping (‘fingerprints’) . . . . . . . . . . . . . . . . Detection reagents for peptides and amino acids . . . . . . . . . Ninhydrin reagent . . . . . . . . . . . . . . . . . . . . . Hypochlorite reagent for detection of ninhydrin-negative peptides Reagents for detection of specific residues in peptides . . . . . .

. . . . . . . . . .

. . . . . . . . . . . . . . . . . . . .

. . . . .

169 169 171 172 173 173 174 176 177 177 179

180 180 181 182 182 183 183 183 183

Appendix I I . . . . . . . . . . . . . . . . . . . . . . List of suppliers . . . . . . . . . . . . . . . . . . . . . . . . . . .

185

. . . . . . . . . . . . . . . . . . . . . . .

189

References

Subject index .

. . . . . . . . . . . . . . . . . . . . .

185

201

This Page Intentionally Left Blank

List of abbreviations

BNPS-skatole 2-(2-nitrophenylsulfenyl)-3-methyl-3-bromoindolenine DFP diisopropylfluorophosphate DTNB 5,S-dithiobis (2-nitrobenzoic acid) DTT dithiothreitol EDTA ethylenediaminetetraacetic acid NEM N-eth ylmaleimide NPS2-nitrophen ylsul fen ylNPS-C1 2-nitrophenylsulfenylchloride NTCB 2-nitro-5-thiocyanobenzoic acid PMB p-mercuribenzoic acid SAEC S-aminoethylcysteine SCMC S-carboxymethylcysteine SMC S-methylcysteine TKL E-N-trimethyl-L-6-hydroxylysine THLP 8-N-trimethyl-L-8-hydroxylysine phosphate 2,4,6-trinitrobenzenesulfonic acid TNBS TNM tetranitromethane TNP2,4, (i-trinitrophenylmolar extinction coefficient in litre mo1-l cm-’ EM change in molar extinction coefficient A% cp phenyl residue 9

Subject index p . 201

CHAPTER 1

Introduction

Chemical modification of proteins is performed for a variety of reasons. In basic research, the two most widely used applications are in amino acid sequence analysis, and in the identification of residues at the catalytic and binding sites of proteins. There is a growing awareness of the fact that a large number of chemical modification reactions on proteins occur in viva, and that such modifications have profound importance in affecting the biological functions of these macromolecules. Indeed, the naturally occurring amino acid derivatives in proteins offer valuable clues to the preparation of chemically related protein derivatives with potentially interesting biological activities. The discussion of analytical procedures (ch. 2) includes methods for the quantitation of both chemically introduced and naturally occurring derivatives of amino acid residues. Since the amino- and carboxylterminal residues of proteins are frequent targets of chemical modification, methods for end-group analysis are presented. Methods for the analysis of amino sugars and carbohydrates are included in recognition of the rapidly growing interest in the modification of glycoproteins (e.g. pituitary glycoprotein hormones, constituents of cell membranes, etc.). Procedures involving group-specific reagents, resulting in the modification of one type of amino acid side-chain (e.g. sulfhydryl), are discussed in ch. 3. Selection of methods for detailed presentation was based either on the high specificity of the modification, or on unique applicability to a general purpose, e.g. the preparation of biologicallyactive protein derivatives radio-labeled to a high specific activity. No 10

Ch. 1

INTRODUCTION

11

complete coverage of all available methodology is intended, and many other experimental procedures may be found in the books and reviews referred to later in this chapter. Group-specific reagents have been used often to achieve a sitespecific modification of a native protein. Since the successful applications are largely the result of chance, experimental procedures are not given (but may be found in the review references listed below). Rather, attention is focussed on a consideration of the factors which lead to the frequently observed unexpected ‘specificity’ of such reagents. A discussion of the various types of ‘affinity’reagents for the binding and catalytic sites of proteins is presented in ch. 5 together with guidelines for the synthesis of typical compounds of each type. In affinity labeling, the information on the binding specificity of a protein for its normal ligand is exploited in the design of a substrate analogue bearing an active group. There is an abundance of recent literature on the chemical modification of proteins. The laboratory-oriented treatment of the subject in Methods in Enzymology, vol. XI (Hirs 1967)and vol. XXVB (Hirs and Timasheff 1972) is particularly comprehensive and valuable. A monograph by Means and Feeney (1971) is also an excellent source of references and methods. Affinity labeling has been reviewed in detail by Singer (1967), Baker (1967), and Shaw (1970a, b), and photoaffinity labeling by Knowles (1972). Other recent reviews include those of Cohen (1968,1970),Freedman (1971), Glazer (1970, 1975), Riordan and Sokolovsky (1971), Spande et al. (1970), Stark (1970),and Vallee and Riordan (1969). The characterization of proteins and their derivatives is outside the scope of this monograph. It is necessary to emphasize, however, that modification of native proteins frequently gives rise to complex mixtures of products. The complexity of the situation is not immediately apparent solely from the stoichiometry of the modification reaction. The now classical case of the reaction of native bovine pancreatic ribonuclease A with iodoacetic acid at pH 5.5 may be cited in this context. One carboxymethyl group is introduced per molecule of Subject index p 201

12

CHEMICAL MODIFICATION OF PROTEINS

ribonuclease, but this is the consequence of the formation of two products in unequal amounts, 1-carboxymethyl-his-12-ribonuclease and 3-carboxymethyl-his-119-ribonuclease (Crestfield et al. 1963). There is a growing list of oligomeric enzymes which display the phenomenon of ‘half-of-the-sites reactivity’ (Levitzki et al. 1971), in which only half of the active sites react with chemically reactive substrate analogues. In such cases, it is necessary to discriminate between the case in which approximately half of the molecules fail to react because of prior denaturation, and cases in which bomfide ‘half-of-the-sitesreactivity’ exists. Finally, treatment with reagents such as tetranitromethane, may lead to products in which the expected tyrosine derivative is absent. The extent of formation of such derivatives is difficult to assess from amino acid analyses. It is appropriate, therefore, to examine protein derivatives, following site-specific modification by electrophoretic and chromatographic methods. Acrylamide gel electrophoresis, both in the presence and absence of sodium dodecylsulfate (Gordon, this series, vol. l), and isoelectric focusing on polyacrylamidegels offer convenient, rapid, and highly sensitive means of examining protein derivatives for homogeneity. Ion-exchange chromatography (Peterson, this series, vol. 2) and gel filtration (Fischer, this series, vol. 1) are powerful tools for the resolution of complex mixtures of derivatives.The followingcompendia summarize much of the relevant literature and references: polyacrylamide gel electrophoretic procedures (Maurer 1971); protein purification (Jakoby 1971);peptide fractionation (Hirs 1967).

CHAPTER 2

Chemical characterization of proteins and their derivatives

2.1. Amino acid analysis with special reference to problem amino acids and common derivatives In many studies of chemical modification of proteins (or peptides) it is essential to determine which amino acid residues have reacted with the reagent being investigated, and the extent of modification. It is not generally appreciated how difficult this can be, particularly for large proteins (or peptides) with multiple residues of the same type. Ideally, the amino acid sequence of the modified protein should be known, and peptides bearing the modified residues should be isolated and characterized. This is greatly facilitated if the modified residues have markers (radioactive, highly fluorescent, etc), which may be detected with high sensitivity. However, in many instances it is desirable as an initial or supplemental method to use amino acid analysis to monitor the results, and even in the ideal situation, peptides with modified groups must be analyzed. Although the precision of analysis with our present technology is often 1 to 3 %, the quantitative release of many amino acids and amino acid derivatives from proteins is often difficult and lowers the overall precision. For example, if constant-boiling HCI (about 5.7 N) is used to hydrolyze a protein in U ~ C U at O 110°C for 24 hr (these conditions are those most commonly used), the amounts of aspartic acid, asparagine, serine, threonine, glutamic acid, glutamine, valine, isoleucine, methionine, tyrosine, tryptophan, cysteine and cystine present in the 13

Subject indexp. 201

14


hydrolysate may be different from the amounts in the protein hydrolyzed. The differences are small for some of these amino acids (histidine and phenylalanine are sometimes also in this category), but large for others. Although some of these amino acids do not have sidechains that are usually modified in proteins, any amino acid can be the NH,-terminal or COOH-terminal residue of a protein and have reactive a-amino or a-carboxyl groups; therefore, quantitative analysis of all amino acids is important. Fortunately, if control samples are hydrolyzed in the same manner as the experimental samples, many problems can be corrected. Some of the problems in preparing and analyzing protein hydrolysates will be discussed below. 2.1.1. Preparation of proteins for hydrolysis It is essential that proteins, evaluated by amino acid analysis, are homogeneous as judged by several physical and chemical criteria (gel electrophoresis, column chromatography, end group analysis, etc.). The purified protein should be separated from substances of low molecular weight, such as inorganic salts, by dialysis against 0.1 M NaCl for at least 1 day (to displace other salts) and then against several changes of deionized water until chloride ions cannot be detekted with AgNO, in the outer solution. Aliquots of the protein solution (or finely dispersed suspension)can then be lyophilized to dryness and used for the determination of ash content (combustion to constant weight in a platinum crucible over an open flame), water content (drying weighed, air-equilibrated samples to constant weight under vacuum at 105OC), nitrogen content (micro-Kjeldahl or Dumas procedures), phosphate determination, carbohydrate analysis (see 82-13),spectroscopy, metal analysis, amino acid analysis, etc. Alternatively, the entire protein preparation may be lyophilized to dryness, extracted with absolute ethanol and ether to remove any lipid material (Light and Smith, 1963), and weighed portions used for the determinations listed above. It is often desirable to separate prosthetic groups from the protein (if possible without modifying amino acid residues) prior to amino acid analysis. If an accurate extinction coefficient for the protein has been determined, it is often convenient to use it as a measure of protein

Ch. 2

CHEMICAL CHARACTERIZATION

15

content rather than the value determined by weighing. After analysis of the protein, the dry weight (less ash) of the aliquot used for analysis should equal the combined weights of the amino acids recovered (calculated from the moles of amino acids found less 1 mole of water per mole of amino acid) plus any other known constituents (phosphate, carbohydrate, prosthetic groups, metal ions, etc.). If this is not the case (within experimental error), thorough examination of the protein for other unknown constituents should be made. Constituents of the protein may be expressed as moles per g (or 100000g) of protein, or if the molecular weight of the protein is known, as moles per mole of protein. In some cases moles per 100 moles of amino acids or molar ratios (based on one amino acid) are used. 2.1.2. Acid hydrolysis of peptides and proteins Aliquots of the soluble or suspended protein, which has been suitably prepared for acid hydrolysis (9 2.1.1), are lyophilized in appropriate glass containers that will withstand the hydrolysis conditions. For this purpose ordinary Pyrex or Kimax test tubes have been used, but many laboratories (including our own) prefer tear-drop glass bulbs with long stems that can be made easily in the laboratory or purchased commercially. The hydrolysis tubes should be cleaned with hot HNO, (concentrated HNO, diluted with an equal volume of water) or other appropriate glass cleaner, and then thoroughly rinsed with deionized water and dried before use. The amount of protein to be hydrolyzed will vary with the size, composition and availability of the protein. However, with most amino acid analyzers currently in use (e.g. the Beckman 120C), it is preferable to have at least 0.01 pmole (but less than 0.60 pmoles) of each amino acid for each column required for analysis, although amounts less than this can be determined less precisely. Therefore, for a protein with a molecular weight of 20000 and having a single residue of histidine, a minimum of 0.4to 0.5 mg of protein should be hydrolyzed for analysis with an analyzer which requires two columns for complete analysis. The lyophilized protein is dissolved or suspended in 1 to 2 ml of 5.7 N HCl, which is usually prepared by glass distillation, repeated Subject index p . 201

16


once or twice to form the constant boiling mixture. One drop of 5 % (w/v) aqueous phenol is usually added for protective purposes. It is necessary to exclude oxygen from the tube during hydrolysis in order to recover maximum yields of the amino acids. Although this can be done by alternatively evacuating and flushing the tube with nitrogen, we have found that gradual evacuation of the tube with the aid of a high vacuum (mechanical)pump and then warming the tube for 30 sec in a 40°C bath with gentle agitation while it is still being evacuated, is an effective method of removing oxygen. Many protein samples will tend to froth under these conditions, and to prevent sample loss it may be necessary to freeze the samples in dry ice-acetone and gently thaw under vacuum, prior to warming. After this procedure the tubes are sealed under vacuum with an oxygen torch and placed in an oven at llO°C for the desired time. The sealing operation may be facilitated by having the hydrolysis tube connected through a short piece of rubber tubing to a stopcock which can be closed and removed from the vacuum system. The standard time ofhydrolysis is 24 hr, but longer times may be necessary for quantitative estimation of some amino acids (see below). After hydrolysis is complete the tubes are cooled, scored and opened, and the acid is removed, either by rotary evaporation or in a heated desiccator over NaOH at 40-50°C.In our laboratory it is standard procedure to examine 1/10 of each hydrolysate by high-voltage paper electrophoresis at pH 1.9 (see Appendix I) prior to analysis to ensure that appropriate amounts are analyzed. Some amino acid derivatives (oxidation products of methionine, methylated lysines, etc.) are also sometimes observed by this procedure. It is usually helpful to add an internal standard, either prior to hydrolysis to reveal hydrolytic losses or prior to analysis to reveal analytical losses. For this purpose we have used norleucine (0.03 pmoles). This amino acid elutes after leucine on the 60 cm column. 2.1.3. Analysis and calculations

The hydrolyzed samples, as prepared above, are dissolved in buffer and analyzed as recommended by the analyzer manufacturer. With a Beckman 120C analyzer we routinely dilute the samples to 1.0 ml with

Ch. 2


17

the recommended pH 2.2 buffer and 0.4 ml is used for each column. The amounts of each amino acid are determined by manual or automatic integration of the area under each amino acid peak. If automatic integration is used, the chart should be examined to be sure that the integration values correspond with the relative positions and peak heights obtained. The calculated amounts of the amino acids (usually in pmoles) can then be converted to moles of amino acid per mole of protein or per 100000 g of protein by using the appropriate correction values for the amount of protein hydrolysate actually used for each column of the analyzer. There has been a recent trend towards the use of high pressure systems and reagents other than ninhydrin to increase sensitivity and speed (e.g. see Moore 1972). Although not yet in general use, such systems hold considerable promise for the future. Expanded scales and one-column systems of analysis have also been used to boost sensitivity. However, in most of these more sensitive systems, the baseline is usually not as straight and free of drift as in the two-column systems, and there are often more problems encountered with interfering substances. At present we would recommend the use of the common commercial analyzers without special modifications (except for the possible inclusion of expanded scale modes) for the greatest accuracy in routine work. However, for situations where samples are obtained in limited amounts or where many samples need to be analyzed quickly, some of the newer systems may well be worthy of consideration. The application of fluorescamine to the quantitative fluorimetric determination of picomole quantities of amino acids, peptides and proteins is of particular interest in this context (Udenfriend et al. 1972). It should be noted that for a protein having 20 residues of one amino acid (e.g. leucine), the precision of the results obtained by analysis of acid hydrolysates of this protein would usually indicate 20 f1residues. As mentioned in the introduction to ch. 2, ifthe procedures described in $2.1.2 and 52.1.3 are used with only one time of hydrolysis, quantitation of at least 13 of the 20 common amino acids may be in error. Special methods that may be utilized to obtain more accurate analyses of these ‘problem amino acids’ and their derivatives are given below. Subiivr index p. 201

18


Many of the details of analysis to be described in this article are based on procedures used in our own laboratory. An attempt will be made to give credit for other procedures. Beckman 120B and 120C analyzers have been used routinely in our own studies.

2.2.

Serine, threonine, tyrosine and derivatives

2.2.1. Analysis

These 3 amino acids are often partially destroyed by acid hydrolysis with the amount of destruction dependent on the time of hydrolysis. With some proteins there is little or no loss of threonine or tyrosine, but serine destruction is rather constant in all proteins. It has been common practice to correct serine values by 10% for each 24 hr of acid hydrolysis, and sometimes threonine and tyrosine values are corrected by 5 % as a first approximation of the extent of destruction. The most accurate method of analysis for serine, threonine and tyrosine in proteins is to hydrolyze the proteins for several different times (e.g. 24, 48 and 72 hr) and extrapolate the values to ‘0’time. In general, the more times that are utilized, the more accurate will be the extrapolation. Extrapolation of the NH, content of these hydrolysates to 0 time will also give an estimate of the number of amides present in the protein, if care has been used to exclude NH, during purification of the protein. The inclusion of phenol in the hydrolysates has generally decreased the rates of destruction, particularly for tyrosine. If serine phosphate (4 2.12.4) or similar derivatives are present, the extrapolated curve will be more complex due to different rates of destruction. Since most derivatives of serine and threonine are not stable to acid hydrolysis, they will not be discussed here. However, those occurring naturally in proteins are described in 4 2.12.4. It should be noted that 0-maleylated derivatives and similar derivatives of serine and threonine have been found as side-products of reactions used to modify other residues in proteins (e.g. ch. 3 ) . 2.2.2. Halogenated and nitrated derivatives of tyrosine A common finding, if proteins are hydrolyzed with undistilled HC1 or

Ch. 2


19

in the absence of phenol, is the conversion of some tyrosine to halogenated derivatives or ‘tyrosine-X’, which is usually the 3-chloroderivative (see Sanger and Thompson 1963). Exposure of certain peptides or peptide hydrolysates to polluted air (‘smog’) has also caused some of the tyrosine to be converted to a substance which elutes in the nitrotyrosine position on the analyzer. It is conceivable that this substance is indeed nitrotyrosine, produced from nitrogen oxides or other pollutants in the air. Halogenated and nitrated derivatives of tyrosine may also be prepared by specific chemical modification reactions (0 3.7). The halogenated and nitrated derivatives of tyrosine may be analyzed by continuing elution of the 60 cm column past the elution of phenylalanine. Elution volumes, determined for one run with a Beckman 120C analyzer, were 204 ml for phenylalanine, 223 ml for nitrotyrosine and 264 ml for tyrosine-X. The various halogenated derivatives of tyrosine would undoubtedly elute from this column at different positions as was found for elution from a 15 an column operated according to Spackman et al. (1958): tyrosine 17.5 ml; 3-chlorotyrosine, 26.0 ml ; 3.5-dichlorotyrosine, 33.3 ml ; 3-bromotyrosine, 31.5 ml ; 3,5-dibromotyrosine, 47.5 ml; 3-iodotyrosine, 42.7 ml ; 3,5-diiodotyrosine, 87.5 ml ; 3-chloro- 5-bromotyrosine, 39.5 ml; lysine, 52 ml; NH,, 77.5 ml (see Sanger and Thompson 1963). In a regular 2-column analysis of a peptide or protein, the presence of peaks on the short column between the positions of phenylalanine and tryptophan is often indicative of halogenated or nitrated derivatives of tyrosine, but further confirmation in other systems should be obtained. We have found that a 60 cm column, eluted with the usual short column (basic) buffer gives excellent separations of many of these derivatives. Sanger and Thompson (1963) also reported the high voltage electrophoretic migrations of the halogenated derivatives at pH 1.85 (2 % formic acid, 8 acetic acid, 1 hr at 4000 volts; this system should be comparable to the pH 1.9electrophoresis described in the Appendix). All migrate slower than tyrosine in this system: tyrosine, 22.5 cm; 3-chlorotyrosine, 19.3 cm ; 3,5-dichlorotyrosine, 16.5 cm; 3-bromotyrosine, 18.0 cm ; 3,5-dibromotyrosine, 14.8 cm ; 3-iodotyrosine, Ssbiecr index p 201

20

CHEMICAL MODIF'ICATION OF PROTEINS

17.0 cm; 3,5-diiodotyrosine, 13.0 cm ; and 3-chloro-5-bromotyrosine, 15.7 cm. The migrations in pH 8.9 electrophoresis are also reported. Nitrotyrosine has a migration of 0.87 compared to 1.00 for tyrosine in pH 1.9 electrophoresis. It should be noted that these derivatives have lower color values thFn tyrosine and that quantitation is generally not good after acid hydrolysis.

2.3. Valine and isoleucine During acid hydrolysis of proteins, valine and isoleucine are often released more slowly than other amino acids due to steric hindrance of hydrolysis by the P-branched sidechains. The Ile-Ile bond is particularly resistant to hydrolysis and is cleaved to the extent of about 50% in 24 hr at 110°C. This could mislead an investigator into the belief that one residue of isoleucine was present in a given peptide rather than two. The Val-Val, Val-Ile and Ile-Val bonds are also slowly hydrolyzed, being 60-75 % cleaved in 24 hr at 110°C. Therefore, if the amounts of valine and isoleucine appear to reach a plateau after 72 hr of hydrolysis, these values may be used; otherwise, hydrolysis for 120 hr may be required to obtain quantitative values. Small amounts of alluisoleucine,which elutes from most analyzer columns just before isoleucine, should be included in the isoleucine determination.

2.4. Aspartic acid, glutamic acid, asparagine and glutamine The amounts of aspartic acid, asparagine, glutamic acid and glutamine in a protein cannot be determined by analysis of acid hydrolysates alone. In order to obtain these values it is necessary to determine the complete sequence or to analyze enzymic hydrolysates which have been prepared by successfully hydrolyzing quantitatively all peptide bonds in the protein (see 0 2.11). Since this is often impractical or difficult to do, it has been customary to report the combined values of aspartic acid and asparagine as Asx (or aspartic acid) and glutamic acid and glutamine as Glx (or glutamic acid). These values can be obtained after acid hydrolysis by the usual procedures outlined above,

Ch. 2

21


since the amides are quantitatively converted to the acids under these conditions. If the amide content has been estimated by extrapolation of the NH, content of the hydrolysates to 0 time (see 4 2.2.2), the combined content of the aspartic acid plus glutamic acid can be estimated, but the amount of either individual component cannot be determined. The combined content of glutamine and asparagine can also be estimated after treatment of the protein with concentrated HC1 in a closed container for 10 days at 37°C (Rees 1946) or with 2 N HCl for a few hours at 100°C (Leach and Parkhill 1955). In this latter procedure, it is necessary to do timed hydrolyses with extrapolation to 0 time for accurate results. In these and other procedures which depend on the production of NH,, controls should be run to determine the amount of free NH, before hydrolysis. A complex procedure for determining the content of asparagine and glutamine separately in proteins has been described (see Chibnall et al. 1958). This procedure makes use of esterification of the protein, reduction of the resulting esters of aspartic acid and glutamic acid with lithium borohydride and hydrolysis of asparagine and glutamine to the acids in which form they are analyzed. Despite the side reactions for which corrections must be made, this method, in conjunction with total enzymic hydrolysis, may be useful to those who must quantitatively estimate these 4 amino acids. It should be noted that if asparagine and glutamine are released from a protein (e.g. by enzymic hydrolysis § 2.1 1.4), methods are available for analyzing them separately (see Benson et al. 1967; Tower 1967).

2.5. Cysteine, cystine, methionine and derivatives These 3 sulfur-containing amino acids and their derivatives are susceptible to oxidation and other destructive reactions. Even when great care has been taken to remove all oxygen from hydrolysis tubes, considerable losses of cysteine and cystine are found after acid hydrolysis, and this usually prevents direct quantitation of these amino acids in proteins. However, total cysteine plus half-cystine content may be determined as cysteic acid after performic acid Subjecl index p . 201

22


oxidation (5 2.5.1) or as S-sulfocysteine. Estimation of cysteine or halfcystine individually may be made by making other derivatives ($3.8), some of which may be analyzed spectrophotometrically and others by amino acid analysis (see below). Since cysteine elutes in the same position as proline, it may be necessary to convert any free cysteine to cystine by air oxidation or to S-sulfocysteine (see below) after acid hydrolysis in order to obtain good results for proline. It should be noted that total cysteine plus half-cystine can also be determined as S-sulfocysteine after acid hydrolysis in the absence of oxygen (Inglis and Liu 1970). This is apparently possible because derivatives of cystine and cysteine that form during acid hydrolysis can be reduced with dithiothreitol to cysteine which is then converted to the S-sulfo derivative by tetrathionate (6 3.8.8). For proteins which contain tryptophan, tetrathionate (Na,S,O,) should also be included in the 6 N HCI prior to hydrolysis. S-sulfocysteineelutes from a 60 cm column of a Beckman 120C analyzer in the same position as cysteic acid (unretarded position) and other substances like 0-phosphoserine, and therefore it is necessary to run control samples to detect interfering substances. S-sulfocysteine has a lower 570 nm/440 nm ratio than cysteic acid and has a color value of 74 % that of aspartic acid under the described conditions of formation. The most widely used procedure for the determination of the half-cystine and cysteine content of proteins is carboxymethylation or carboxyethylation. If appropriate precautions have been taken in the preparation of a protein, and if oxygen is completely removed before hydrolysis, methionine will usually be recovered from acid hydrolysates in yields greater than 95 %. However, in some proteins (particularly those that are chemically modified) and in many peptides the methionine may be at least partially oxidized to the sulfoxide or sulfone forms, and even though these may be analyzed with amino acid analyzers (see below), the total yield of methionine (and oxidized products) is usually somewhat low. A good check on total methionine content in a peptide or protein is obtained by analyzing for methionine sulfone after performic acid oxidation, since methionine and its sulfoxides are quantitatively converted to the sulfone by this procedure.

Ch. 2


23

2.5.1. Performic acid oxidation (for analytical work) After oxidation of unmodified proteins with performic acid (0 3.8.1) the methionine will all be present as methionine sulfone, and the cysteine and cystine will be present as cysteic acid. The method for oxidizing proteins with performic acid on a preparative scale ($ 3.8.1) has been modified for analytical studies (Moore 1963),to allow rapid destruction of excess performic acid. However, tryptophan is still destroyed by the performic acid in this modified procedure, and quantitation of tyrosine may be complicated by the formation of halogenated derivatives (9 2.2.3). The performic acid is prepared in the usual manner by allowing a mixture of 1.0 ml of 30% (w/v) H 2 0 2and 9.0 ml of 88% (w/v) formic acid to stand at room temperature for 1 hr, after which it is cooled to 0°C. A known amount of protein, usually the amount used for routine hydrolysates, is dissolved in 1 to 2 ml of the reagent in a cooled hydrolysis tube. If the protein does not dissolve readily in the cold reagent, it can usually be dissolved first in 0.1 ml of 100% formic acid at room temperature. The mixture is kept at 0°C for 4 hr after which performic acid is destroyed by the addition of0.15 ml of cold 48 % (w/v) HBr per ml of performic acid reagent used. The bromine which forms, as well as the formic acid solution, can best be removed from the hydrolysis tube by rotary evaporation under high vacuum (mechanical pump) at 40°C with NaOH pellets in the condenser trap; the condenser should be cooled with dry ice in ethanol. Acid hydrolysis of the oxidized sample and analysis of the resulting hydrolysates are performed in the usual manner (92.1.2). 2.5.2. Cysteic acid Cysteic acid is the major product of performic acid oxidation of cysteine and cystine in proteins, and is usually produced in yields of more than 90%. Cysteic acid is not retarded by the resins generally used in amino acid analyzers and therefore elutes at the breakthrough volume (about 42% of the elution volume of aspartic acid). However, since other substances such as 0-phosphoserine can also elute in this position, it is essential to analyze hydrolysates made both before and Subjccr index p. 201

24


after performic acid oxidation of the protein to determine if interfering substances are present. Cysteic acid has a color value of 101% that of aspartic acid (Spackman et al. 1958),and in electrophoresisat pH 1.9 it has a mobility of 0.47 towards the anode compared to 1.00 towards the cathode for aspartic acid. 2.5.3. S-Carboxymethylcysteine (SCMC) and S-carboxyethylcysteine S-Carboxymethylcysteine is prepared by alkylation of cysteine with iodoacetic acid (43.8.2). Iodoacetamide gives the amide derivative, which is converted to SCMC by acid hydrolysis. Since SCMC (and other derivatives of cysteine) is especially susceptible to oxidative and other destructive reactions, it may be helpful in some instances, to include one drop of 5 % mercaptoacetic acid (or some other reducing agent) in addition to the phenol in the 6 N HCl prior to hydrolysis. However, blanks should be run (no protein) to correct for any interfering ninhydrin-positive substances derived from the reducing agent. SCMC elutes from the 60 cmcolumn of a Beckman 12OC analyzer at a volume of 46 ml compared to 52 ml for aspartic acid. The color value for SCMC is 92% that of aspartic acid (Spackman et al. 1958). It has recently been observed that SCMC can cyclize, particularly at acid pH (optimally at pH 3) to give a thiazine derivative which is ninhydrin-negative(Bradburyand Smyth 1973)(4 3.8.3). For this reason it may often be preferable to make the S-carboxyethylcysteinederivative in proteins (4 3.8.3) as it does not readily undergo this type of reaction (Bradbury and Smyth 1973). S-carboxyethylcysteine elutes from a 60 cm column ofa Beckman 120Canalyzer between serineand glutamic acid; its color value should be similar to that of SCMC. 2.5.4. S-Aminoethylcysteine (SAEC) This positively-charged derivative is prepared from cysteine and ethylenimine (4 3.8.4) and is often used to introduce additional sites of tryptic hydrolysis into proteins. SAEC has a mobility of 2.07 compared to 1.00 for aspartic in pH 1.9 electrophoresis, and elutes from a 20 cm (basic) column of a Beckman analyzer at 82 ml compared to 74 ml for lysine and 96 ml for histidine (Schroeder et al. 1967).SAEC has a color

Ch. 2


25

value of 92 % that of lysine and 88 % that of leucine, if losses due to acid hydrolysis are included (Hofmann 1964).

2.5.5. S-Methylcysteine (SM C ) S-Methylcysteine, which is a neutral derivative produced by alkylating cysteine with methyl-p-nitrobenzenesulfonate (9 3.8.6) has a mobility of 1.04 compared to 1.00 for aspartic acid in pH 1.9 electrophoresis and elutes from the 60 cm column at the trailing edge of the proline peak (Heinrikson 1971).This often prevents accurate estimation of proline, but if the proline content is low, there is little problem in quantitating the SMC derivative. If better separations are desired, the investigator suggests the use ofa 150 cm column, but it also seems likely that changes in ionic strength or pH of the buffer or changes in the temperature of the column would also allow better separations with the 60 cm column. The color values for SMC are 98 % (unhydrolyzed) or 89 % (including correction values for losses during acid hydrolysis) that of alanine. A decomposition product (possibly the sulfone) is formed during the first 22 hours of acid hydrolysis, but not thereafter, and elutes at the position of the leading edge of aspartic acid. Perhaps the amount of this decomposition product could be reduced if phenol or a reducing agent is included during hydrolysis (§ 2.1.2). 2.5.6. Thialaminine Thialaminine, or the basic ethyltrimethylammonium derivative of cysteine, is prepared from cysteine and (2-bromoethy1)trimethylammonium bromide (9 3.8.5). Thialaminine is stable to acid hydrolysis for periods of at least 72 hr at 106"C, elutes from a 10 cm column of PA-35 resin on the Beckman analyzer between lysine and histidine and has a color value 86% that of lysine (Itano and Robinson 1972). Another basic derivative of cysteine, S-(4-pyridylethyl)-L-cysteine has also been characterized. It elutes from the short column of the analyzer just before arginine and has a color value of 102% that of leucine (Friedman et al. 1970).

2.5.7. S-Succinylcysteine (S-l,2-dicarboxyethyl-L-cysteine) S-Succinylcysteine is the product formed from the reaction of cysteinyl Tiihierr inderp 2111

26


residues in proteins with maleic anhydride (4 3.1.3.2).Proteins in which the cysteinyl residues have been titrated with N-ethylmaleimide (9 3.8.9.2) also give S-succinylcysteine and ethylamine in equal amounts after acid hydrolysis (Smyth et al. 1964). The best yields are obtained after 72 hr of hydrolysis (Smyth et al. 1964) and after careful removal of oxygen (yields as high as 95 %; Guidotti and Konigsberg 1964). If the ratio of ethylamine to S-succinylcysteine is greater than 1 after acid hydrolysis ofthe protein, either the protein was not completely separated from the reagent (N-ethylmaleimide) which also gives ethylamine on hydrolysis, or other groups may also have reacted. The reagent has been shown to react with the nitrogens of lysyl sidechains, histidyl rings and a-amino groups (Holbrook and Jeckel 1969). S-Succinylcysteine elutes from the long column before both S-carboxymethylcysteine and aspartic acid; with the 150 cm column system of Spackman et al. (1958), Smyth et al. (1964) obtained an elution volume of 108 ml for S-succinylcysteine compared to 131 ml for aspartic acid. The color value of S-succinylcysteine is 109% that of aspartic acid. Ethylamine elutes from a 15 cm short column at 117 ml compared to 135 ml for arginine; the color value of ethylamine was determined as 8.3 (which is 42% that of aspartic acid on the other column). 2.5.8. Methionine suljone Methionine sulfone is a product of oxidation of proteins with performic acid (9 2.5.1) or other strong oxidizing agents. Milder oxidation of methionine gives the sulfoxide derivatives (9 2.5.9.),which are also converted to the sulfone by performic acid. In general, quantitative yields of methionine sulfone are obtained by oxidation with performic acid. Methionine sulfone is stable to acid hydrolysis and elutes from most analyzer columns immediately after aspartic acid. Since the elution position of methionine sulfone is more sensitive to temperature changes than that of aspartic acid, these two amino acids may not be well resolved in some systems. However, temperature adjustments of the columns usually allow better separations to be obtained with the

Ch. 2


21

60 cm columns; alternatively, the 150 cm system described by Moore et al. (1958)and Spackman et al. (1958)can be used for better resolution. The color value for methionine sulfone is the same as for aspartic acid (Spackman et al. 1958).

2.5.9. Methionine sulfoxides These derivatives of methionine (isomers) may be obtained by mild oxidation of proteins or peptides with reagents like H,O, under controlled conditions (see also Q 3.5). They are also found in peptides and proteins (particularly if denatured) which have been exposed to air for prolonged periods. Since the usual conditions of acid hydrolysis convert much of the sulfoxide to methionine (see Ray and Koshland 1960), the extent of methionine oxidation may not be recognized, unless other analytical procedures are utilized (see below). In our own laboratory we have studied peptides in which the methionine sulfoxides are recovered from acid hydrolysates in 5-30 % yield, and the combined recoveries of methionine and methionine sulfoxides are usually in the range of 85-95 % Methionine sulfoxides elute from most analyzer columns just prior to aspartic acid; there are two separate peaks observed with the 150 cm column system of Spackman et al. (1958) and sometimes with the 60 cm column systems in current use. The color value for methionine sulfoxides is 99 % that of aspartic acid (Spackman et al. 1958). At present there are three methods which can be used for analyzing methionine sulfoxides quantitatively in peptides and proteins : analysis after complete enzymic hydrolysis (see 9 2.1 I), analysis after alkaline hydrolysis (see below) and analysis after converting the nonoxidized methionine to the carboxymethyl sulfonium derivative, followed by oxidation of the methionine sulfoxides to the sulfone with performic acid (see 5 2.5.10). The enzymic method will usually give excellent results with peptides (less than about 30 residues) if care is taken to prevent further oxidation during the reaction and handling of the hydrolysate (a nitrogen barrier is recommended). However, for larger peptides and proteins, incomplete enzymic hydrolysis may give erroneous results, and one of the other two methods is recommended. Svbpci m C . y p . 201

28


Methionine sulfoxide may be released essentially quantitatively by hydrolysis with 15% (- 4.4 M) NaOH (see Neumann et al. 1962).The procedure described by these investigators calls for hydrolysis of 5 mg of protein in special vapor-tight, screw-cap Telfon vials, but it is possible to hydrolyze 1 mg or less if the hydrolysis is carried out in hydrolysis tubes with small volumes and containing a polypropylene liner to hold the sample in a manner similar to that described by Hugh and Moore (1972). Convenient liners for these small amounts of protein and NaOH are small conical plastic tubes (such as the micro sample tube, Dynalab Corp., Rochester, N.Y.). The dried sample is dissolved in 0.2 ml of 15 % NaOH (freshlymade from 50% NaOH each time) for each mg of protein, and after evacuating the tube cautiously to remove bubbles without loss of sample, the tube is sealed and placed in an oven at 110°Cfor 16 hr. Alternatively, the tube may be successively evacuated and flushed with nitrogen several times before sealing. After hydrolysis the tube is cooled before opening, and the solution is acidified with 0.19 ml of 6 N HCl and 0.1 ml of 1.1 M citric acid (containing 0.05 ml of thiodiglycol per ml) per 0.2 ml of 15% NaOH used, to give a final pH lower than 2.2. The solution is diluted to 1.0 ml for each mg of protein hydrolyzed, and aliquots of 0.5 to 1.0 ml are analyzed (e.g. on the 60 cm column of a Beckman 120C analyzer). The 150 cm column system described by Moore et al. (1958) may also be used. The method described above gives direct analysis of methionine sulfoxide content in proteins. Another method makes use of carboxymethylation of methionine at acid pH to give the carboxymethylsulfonium derivative (4 3.5). Methionine sulfoxide, which is not affected by the carboxymethylation reaction, is then oxidized to the sulfone which is stable to acid hydrolysis and can easily be quantitated. This is possible because methionine carboxymethyl-sulfonium salts are not affected by performic acid oxidation, although they are degraded by acid hydrolysis. Therefore, the methionine sulfone content is equal to the methionine sulfoxide content plus any sulfone that may have been initially present (shown by analysis before oxidation).

Ch. 2


29

2.5.10. Methionine carboxymethylsulfonium salts These derivatives of methionine (isomers) are prepared by treating proteins with iodoacetic acid; the reaction is most specific for methionine at acid pH ($ 3.5). These derivatives are not affected by performic acid oxidation (see under methionine sulfoxide), but are degraded by acid hydrolysis to give methionine, carboxymethylhomocysteine, homoserine and homoserine lactone (Gundlach et al. 1959). Carboxymethylhomocysteine elutes from most analyzer columns just after proline, and the other two derivatives are discussed below. It should be possible to quantitate the carboxymethylsulfonium derivative after complete enzymic hydrolysis (see $ 2.11), since it has been shown that they elute as two peaks just prior to aspartic acid (Gundlach et al. 1959). The amount of carboxymethylsulfonium derivative present can best be quantitated by oxidizing the noncarboxymethylated methionine to methionine sulfone ( 5 2.5.9) and subtracting the amount of methionine sulfone from the total methionine content of the protein to obtain the amount of methionine carboxymethylsulfonium salts. 2.5.1I . Homoserine and homoserine lactone Treatment of proteins with cyanogen bromide results in cleavage of the peptide chain COOH-terminal to methionyl residues with concomitant conversion of the methionine to homoserine lactone which is in equilibrium with homoserine. In certain instances (e.g. -Met-Seror -Met-Thr-) some of the methionine is converted to homoserine without peptide bond cleavage (see Schroeder et al. 1969).Homoserine and its lactone are also products of the breakdown of the carboxymethylsulfonium salts of methionine (4 2.5.10). Homoserine elutes from the 60 cm column of a Beckman 120C analyzer just prior to glutamic acid and has a color value 95 7; that of leucine (Hofmann 1964).Better separation of homoserine and glutamic acid can be achieved by lowering the pH of the eluting buffer to 3.20 (Ambler 1965) or to 2.80 (Schroeder et al. 1967). Homoserine has a mobility of 1.37 compared to 1.00 for aspartic acid in pH 1.9 electroSubierr mdcv p . 201 phoresis.

30


Homoserine lactone elutes from the short column (basic)immediately after NH, and has a color value of 57 % that of lysine (Hofmann 1964). The mobility of the lactone in electrophoresis at pH 1.9 is 2.45 compared to 1.00 for aspartic acid. The lactone gives a yellow-brown color with ninhydrin (homoserine gives a blue color). Solutions of homoserine and its lactone slowly reach equilibrium under most conditions at room temperature. As a result, samples of these derivatives prepared for analysis often gradually change in concentration even during the chromatographic steps, and color values usually reflect such losses. Acidic conditions favor lactone formation, and trifluoroacetic acid at 20°C for 1 hr can be used to convert essentially all of the homoserine (free or in peptides) to the lactone form (Ambler 1965). Alkaline conditions favor homoserine formation, but may often be too harsh for many peptide procedures (see Ambler 1965).

2.6. Tryptophan and derivatives Little or no tryptophan will usually be found in protein hydrolysates prepared with acid in the usual manner. Mercaptans added to the 6 N HCl usually increase the yields of tryptophan (Matsubara and Sasaki 1969), and substitution of p-toluenesulfonic acid (with added 3-(2-aminoethyl) indole) in place of the HC1 has been used to good advantage (Liu and Chang 1971). At the present time the hydrolytic methods of choice for tryptophan analyses appear to be with methanesulfonic acid in the presence of added 3-(2-aminoethyl)indole (see Moore 1972) or with NaOH in the presence of starch (Hugli and Moore 1972). Both methods will be described below, but it should be noted here that hydrolysis with NaOH has the advantage that carbohydrate in the protein sample does not affect the yields of tryptophan as it does in the acid hydrolysates. Other methods of tryptophan analysis involving spectrophotometry (Goodwin and Morton 1946 ;Edelhock 1967),titration with N-bromosuccinimide (Spande and Witkop 1967), or the formation of colored derivatives (Spies and Chambers 1948 ; Barman and Koshland 1967 ;

Ch. 2


31

Spies 1967; Scoffone et al. 1968)may be used to give initial estimates or verification of tryptophan content. It should also be possible to obtain good tryptophan analyses after complete enzymic hydrolysis (§ 2.1 1). Hydrolysis of proteins with methanesulfonic acid is carried out in the usual type of hydrolysis tube with 4 N methanesulfonic acid, containing 0.2% 3-(2-aminoethyl) indole, at 115OC for 24 hr (Liu; see Moore 1972). The acid mixture is available in sealed vials from Pierce Chemical Corp. The usual precautions to remove oxygen should be followed. Since the acid is not volatile, it is necessary partially to neutralize the cooled hydrolysate with an equal volume of 3.5 N NaOH and dilute with water to a volume 5 times the volume of acid used for hydrolysis. The procedure recommends the use of 1 ml of the methanesulfonic acid mixture (for about 2-5 mg of protein), which would result in 5 ml of diluted hydrolysate. Since 1-ml aliquots are analyzed (see below), only 20 % of the protein hydrolyzed is utilized for each analysis. However, where the amount of protein available for hydrolysis is limited, it should be possible to scale down the amounts of materials used to levels such as 0.2 ml of the acid mixture for 0.5 mg of protein if appropriately smaller hydrolysis tubes are used. Aliquots of 1 ml of the hydrolysate are used for analysis with the short (basic) column of a Beckman analyzer ; elutions from a 10 an column of PA-35 resin being: tryptophan, 30 ml; lysine, 37 ml; histidine, 45 ml; and arginine, 82 ml. However, mannosamine and galactosamine coelute with tryptophan in this system and if they are present (determined by extending a 60 cm column run for 60 ml past phenylalanine; see also 3 2.13.1), a 20 cm column of the same resin at S ° C , eluted at 50 ml per hr with 33 ml of 0.2 N sodium citrate buffer at pH 4.25 and then with 230 ml of 0.35 N buffer at pH 5.28, should be used (Liu and Chang 1971). The elution volumes in this system are: tyrosine, 42 ml; phenylalanine, 46 ml; glucosamine, 82 ml; mannosamine and galactosamine, 90 ml; tryptophan, 98 ml; lysine, 125 ml; histidine, 136 ml; arginine, 250 ml. The color value obtained for tryptophan in this system is 78 % that of lysine. With proteins hydrolyzed with HCl (or in the presence of some HCl), an acid decomposition product may be observed at 94 ml in this system. Srrhject index p 201

32

CHEMICAL MODIPiCATION OF PROTEINS

Hydrolysis of proteins (1 to 5 mg) with 4.2 N NaOH (0.6 ml) in the presence of partially hydrolyzed starch (25 mg) has been reported by Hugh and Moore (1972)to give yields of tryptophan greater than 97 % for several proteins. These investigators used the bottom portions of polypropylene centrifuge tubes as liners in regular glass hydrolysis tubes to avoid silicate formation. It is also possible to hydrolyze smaller amounts of protein in smaller volumes if small conical plastic tubes (such as the micro sample tube, Dynalab Corp., N.Y.) are used in smaller hydrolysis tubes. Care should be taken not to melt the plastic liners when sealing the tubes. In practice, the protein (in 0.1 ml of 0.005 N HCl or NaOH is added to a 10.9x 50 mm liner inside a 16x 150 mm thin-walled Pyrex test tube, and 0.5 ml of 5 N NaOH (freshly prepared from 50% NaOH) is added after adding 25 mg of partially hydrolyzed starch. The addition of l-octanol (5 p1 of a 1% solution in toluene) aids in preventing frothing during evacuation. A constriction (to about 2 mm) is made in the glass tube near the middle, and the lower portion is cooled in a dry ice acetone bath for 1.5 min (do not freeze). The tube is evacuated with a high vacuum pump by repeatedly opening and closing a stopcock while tapping the tube to dislodge air bubbles. When the pressure has stabilized between 25 and 50 microns of mercury, the tubes are sealed and placed in an oven for hydrolysis at 110°C for 16 hr. If Val-Trp or Ile-Trp bonds are present, (or if it is not known if they are present) it is necessary to hydrolyze one sample for at least 98 hr at llO°C or 48 hr at 135OC, since these bonds are only slowly hydrolyzed. After carefully cooling the hydrolysates (to prevent loss of sample), the tubes are opened and 0.5 ml of 0.2 M citrate buffer at pH 4.25 (without Brij 35) is added and mixed. The solution is then transferred with several rinses of buffer to a chilled (dry ice) 2.5 ml (or 5 ml) volumetric flask, containing 0.42 ml of 6 N HC1, and diluted to volume with the pH 4.25 buffer. The precautions of using pH 4.25 buffer and the chilled HCl are for the purpose of avoiding losses of tryptophan, known to occur in most acidic solutions. Aliquots of 1.0 ml of the diluted hydrolysates are analyzed on the short column (Beckman analyzer), eluted at 50ml per hour with a buffer prepared by diluting the usual pH 5.3 buffer to give a final Na'

Ch. 2


33

concentration of 0.21 M (the final pH is 5.4) (Hugli and Moore 1972). This buffer separates tryptophan from lysinoalanine, which is a product formed from lysine and dehydroalanine during alkaline hydrolysis of many proteins (Bohak 1964); tryptophan, 37 ml; lysinoalanine, 55 ml; lysine, 68 ml; and histidine, 79 ml. These results are for a 0.9 x 8 cm column of Beckman PA-35 resin, but better results are often obtained on a 12 cm column, which allows more time for the baseline to be established after the neutral and acidic amino acids are eluted. Any ornithine formed from arginine during alkaline hydrolysis will elute just prior to lysine. Hexosamines (see Q 2.13) are destroyed by alkaline hydrolysis, and neither they nor hydroxylysine (see Q 2.12.1.3) interfere with the tryptophan analysis. The color value for tryptophan in this particular system was determined to be 79% that of lysine (Hugli and Moore 1972). 1-Formyltryptophan. This derivative of tryptophan, which is formed in proteins by formylation in anhydrous formic acid saturated with gaseous HCl (Q 3.6.1), is not stable to acid or alkaline hydrolysis, but could be obtained after complete enzymic hydrolysis of derivatized proteins (Q 2.11). With a 0.9 x 7 cm short (basic) column on a Beckman 120B Analyzer, I-formyltryptophan elutes in 53 ml compared to 37 ml for tryptophan (Previero et al. 1967a). In paper chromatography with I-butano1:acetic acid:water (4: 1 : 5, v/v) tryptophan has an R, of 0.50 and 1-formyltryptophan has an R, of 0.58. The amount of derivative formed in proteins may be determined by lyophilization of the solution after the formylation reaction has ceased (usually by 40 min as measured by change in absorbance at 298 nm) and measuring the absorbance at 298 nm of an aliquot of the protein dissolved in 8 M urea at pH 4. The molar extinction coefficient of 1-formyltryptophan has been calculated as 4880 at 298 nm, and the change in absorbance at 298 nm can be used as a measure of tryptophan content in proteins where this is not known (Previero et al. 1967b). Other derivatives of tryptophan which may also be obtained after enzymic hydrolysis of derivatized proteins are those obtained by reaction with sulfenyl halides (Scoffone et al. 1968)and 2-hydroxynitrobenzyl bromide (Barman and Koshland 1967) ( Q 3.6.2). Subject indexp. 201

34


2.7. Lysine derivatives Analyses for lysine after acid hydrolysis of proteins are usually quantitative and reliable, unless the lysines have been exposed to certain types of reagents (e.g. cyanate, nitrous acid). In a few proteins some lysines are E-N-methylated (see (j 2.12.1.2); the methyllysines are stable to acid hydrolysis and emerge from many analyzer columns as a shoulder on the trailing edge of lysine, thus making lysine quantitation less precise (see 9: 2.12.1). Methyllysine may also be formed in proteins by chemical modification (§ 3.1.1.3). Many derivatives of lysine are not stable to acid hydrolysis (e.g. c-N-acetyllysine) and will not be discussed here, but those which occur naturally in proteins are discussed in 9: 2.12.1. 2.7.1. c-N-Carboxymethyllysines Both the E-N-monocarboxymethyl and the E-N-dicarboxymethyl derivatives of lysine are formed by reaction of proteins with iodoacetate at alkaline pH values at which lysyl residues are at least partially unprotonated ((j 3.8.2). Both derivatives are stable to acid hydrolysis under the usual conditions. Both the mono- and di-derivatives are analyzed with the longer column used for analysis of the acidic and neutral amino acids. In the original studies of Gundlach et al. (1959), the 150 cm column was used : c-N-dicarboxymethyllysine, 90 ml ; aspartic acid, 122 ml; E-N-monocarboxymethyllysine, 335 ml ;and methionine, 341 ml. The dicarboxymethyl derivative elutes after cysteic acid but before methionine sulfoxides. To our knowledge other analytical systems have not been used, but the 60 cm column may not give adequate separations, especially for the mono-derivative, since it elutes prior to methionine which is already close to the buffer change peak. It seems likely that by appropriate manipulation of times of buffer change, the pH or ionic strength of the buffer, or the temperature of the column, good separations could be obtained with the 60 cm column. 2.7.2. Homoaryinine Homoarginine is the product of guanidination of lysyl residues with

Ch. 2


35

either 0-methylisourea or reagents such as 1-guanyl-3,5-dimethylpyrazole nitrate (see $3.1.1.2).Like arginine, homoarginine is stable to acid hydrolysis, but not to alkaline hydrolysis. Homoarginine elutes from the short (basic) column of the analyzer after arginine; e.g. with one system homoarginine has an elution volume 1.33 times that of arginine (Habeeb 1960). The color constant should be similar to that of arginine. The migration of homoarginine in a number of chromatographic systems has been described (Bell 1962). Homoarginine has a mobility of 2.20 compared to 1.00 for aspartic acid in pH 1.9 electrophoresis. 2.7.3. Deamination products of lysine Treatment of proteins with nitrous acid can lead to deamination of the lysyl residues as well as of the NH,-terminal residues. Kurosky and Hoffmann (1972) have identified three products of lysine deamination which elute from the 60 cm column of a Beckman 120C Analyzer at (1) the position of glutamic acid, (2) 7 min before glycine and (3) 10 min before tyrosine. Products (2) and (3), which are the major products formed, are believed to be e-hydroxynorleucine and its lactone respectively. The most accurate method for analyzing for lysine deamination is to measure the loss of lysine rather than to measure the three derivatives. 2.7.4. Homocitrultine The &-aminogroups of lysyl residues in proteins may be carbamylated by cyanate to give homocitrulline residues ( Q 3.1.2.1).It should be noted that some carbamylation of proteins can occur in urea solutions that have not been deionized to remove cyanate. The extent of modification is often difficult to assess, however, since acid hydrolysis of fully carbamylated proteins under the usual conditions gives homocitrulline plus 17-33% free lysine (Stark and Smyth 1963). Although the lysine recovery (in %) is variable, depending on the protein studied, for a given protein the lysine recovery is apparently constant. If accurate quantitation of homocitrulline content of a protein is necessary, complete enzymic hydrolysis (6 2.1 1) might be tried. Sirbpw mdeu p 201

36


Homocitrulline elutes from most analyzer columns just prior to valine. Alternatively, the unreacted lysines can be converted under denaturing conditions to other acid-stable derivatives (homoarginine, methyllysines, carboxymethyllysines) which can be quantitated after acid hydrolysis (above and Q 2.12). The homocitrulline content can then be assumed to be the difference between the total lysine content and the content of the acid-stable derivative. Alkaline hydrolysis of carbamylated proteins gives quantitative conversion of homocitrulline to lysine (Stark and Smyth 1963), and this can be used to check the homocitrulline content of the protein in which the other lysines have been converted to another derivative that does not give lysine by alkaline hydrolysis (see also 5 3.1.2.1.).

2.8. Histidine and derivatives Although analyses for histidine after acid hydrolysis are usually satisfactory, the values found may be slightly low, particularly after chemical modification of other residues. This has often led to estimates of the number of residues in a protein that are too high, when histidine has been used to calculate a one-residue value. Naturally-occurring derivatives of histidine (methylhistidines and phosphohistidines) are described in 0 2.12.2. The carboxymethylhistidines are the best characterized derivatives of histidine formed in proteins in uitro. 2.8.1. Carboxymethylhistidines Nitrogens 1 and 3 in the imidazole ring of histidyl residues in proteins may be alkylated with iodoacetic acid (generally in a much slower reaction than alkylation of cysteinyl residues) to give three carboxymethyl derivatives: 1-carboxymethylhistidine, 3-carboxymethylhistidine and 1,3-dicarboxymethylhistidine (Q 3.4). In general, the 3-carboxymethyl derivative is formed most rapidly. These derivatives are stable to acid hydrolysis under the usual conditions (but excess reagent must be removed) and may be analyzed on the long column of most analyzers as described below.

Ch. 2


37

Using the 150 cm column of Spackman et al. (1958), Crestfield et al. (1963) showed that 1,3-dicarboxymethylhistidineelutes about 55 ml before aspartic acid (124 ml), I-carboxymethylhistidine elutes about 9 ml after glutamic acid (177 ml) and 6 ml before proline (193 ml), and 3-carboxymethylhistidine elutes about 16 ml after alanine (264 ml). The color values of the dicarboxymethyl- and monocarboxymethylderivatives were determined as 96 %and 99 % that of serine respectively, and 97 % and 100% that of glycine respectively. The dicarboxymethyl derivative has been characterized in electrophoretic and chromatographic systems, and all three derivatives were separated by ion exchange chromatography in pyridine-formic acid buffer at pH 3.25 (Banaszak and Gurd 1964). 2.8.2. Other derivatives of histidine Many derivatives of histidine are not stable to acid hydrolysis and are not discussed here (however, see 8 2.12.2 for those that occur naturally in proteins). Brief mention should be made of the iodination of histidyl residues by HOI (0 3.7.2). The mono- and diiodohistidines can be identified and distinguished from the iodinated tyrosines by high voltage paper electrophore&s in 1 M formic acid (Roholt and Pressman 1972) after complete enzymic hydrolysis of the protein or peptide (§ 2.11; Roholt and Pressman 1972). Quantitation and identification are facilitated by the use of radiolabeled reagent.

2.9. Arginine and derivatives There is usually little difficulty in obtaining satisfactory analyses for the arginine content of proteins hydrolyzed with acid under the usual conditions. There are few derivatives of arginine that are included in the scope of this treatise. Derivatives of arginine obtained by treating proteins with diketones appear to be the best described to date (see 4 3.2), but acid hydrolysis destroys these derivatives with 12% regeneration of arginine noted in at least one instance (Yankeelov 1970). Subjrcr inde\ p . 201

38


2.10. Other amino acids As stated previously, any amino acid may be the NH,-terminal or COOH-terminal residue of a protein and be subject to destructive reactions during chemical modification studies. In general in addition to some of the amino acids described above, analyses for glycine, alanine and leucine are usually very statisfactory. Values for phenylalanine are sometimes slightly low, and due to its lower color value, proline gives less precise results than other amino acids, particularly for proteins with low proline contents.

2.11. Complete enzymic hydrolysis of proteins andpeptides Because of the lability of several amino acids and many of their derivatives under conditions generally required for quantitative hydrolysis of peptide bonds, a number of attempts have been made to use enzymes under mild conditions to hydrolyze proteins and peptides completely. In theory this seems possible, but in practice it has been difficult to achieve. In retrospect some of the major difficulties encountered in attempts to hydrolyze proteins completely with enzymes are: (I) inability to maintain proteins (or regions of proteins) in a completely denatured state under the mild conditions employed for enzymic hydrolysis (it has been demonstrated that many native proteins are highly resistant to proteolysis) ;(2) the formation of peptide aggregates which are resistant to proteolysis; (3) the presence of certain bonds (e.g. those involving proline) or sequences (e.g. several acidic residues in a cluster) which are poorly hydrolyzed by most enzymes; (4) the presence of blocking groups (e.g. cr-N-acetyl) and large prosthetic groups (e.g. covalently bound heme in cytochrome c) which restrict proteolysis in certain regions; (5) the formation of cr, P- or P-peptide bonds from aspartyl residues (these are not hydrolyzed by most proteases); and (6) the cyclization of glutamine which often causes low yields of glutamine in hydrolysates ; if glutamine occurs as the NH,-terminal residue of peptides, it is difficult to hydrolyze these peptides completely. In addition, it should be noted that proteins containing disulfide bonds

Ch.2


39

are often not hydrolyzed completely by enzymes unless the disulfide bonds are reduced (and preferably alkylated, 5 3.8.2). In general each protein must be examined as an individual case, since a procedure successful for one protein may not work well for another. Peptides are generally hydrolyzed more successfully by enzymes than are proteins. Some recommended procedures that may be tried are described below. Regardless of the method used, the important criterion for success is the quantitative recovery of those amino acids or derivatives in which one is interested. For example, if one has made a radiolabeled derivative of cysteine in a protein, all of the radioactivity must be demonstrated to elute in the position of this derivative after complete enzymic hydrolysis; if it does not, it is not certain whether the enzymic hydrolysis was incomplete or if other residues have also been modified. Although any of several combinations of proteases can be used, ideally, one or more non-specific endopeptidases should be used first to convert the protein into many small peptides. These small peptides can then be degraded to amino acids by aminopeptidases and prolidase (hydrolyzes X-Pro bonds). Sometimes, carboxypeptidases are also used. Although leucine aminopeptidase has been used as the aminopeptidase (see Hill and Schmidt 1962), it may be preferable to use aminopeptidase M (Rohm and Haas, supplied by Henley and Co. of N.Y.), since this enzyme removes most residues at acceptable rates. Leucine aminopeptidase removes hydrophobic residues most rapidly, whereas some other residues are removed very slowly. Most procedures should probably include the use of prolidase (Miles) since many aminopeptidases do not cleave X-Pro bonds at appreciable rates. If it is found that proline is not released quantitatively by these procedures, the use of citrus leaf carboxypeptidase C (Rohm and Haas) can be tried after the initial endopeptidase hydrolysis and before the addition of aminopeptidase M and prolidase. Carboxypeptidase C (also yeast carboxypeptidase Y - see Hayashi et al. 1973) hydrolyzes proline bonds (as well as all others), but if proline is at or adjacent to the NH2 terminus of a peptide, it would probably not be released. In all procedures a control consisting of the enzymes only should be run in parallel with the hydrolyzed sample, and corrections should be made for any amino acids found by analysis of the control. w ~ J/I ~c A ~ :01 ~

40


2.1 1 .I. Procedure 1 : pepsin, subtilisin, aminopeptidase M, prolidase

This method may often be useful with proteins that can be maintained in a. largely denatured state at acid pH. The protein (usually about 10 mg per ml) may be denatured in solvents like 6 M guanidinium chloride or 8 M freshly deionised urea and dialyzed exhaustively against 5 % formic acid at room temperature at about pH 2. If cysteine or cystine residues are present, these should be reduced and alkylated (9 3.8.2) prior to enzymic hydrolysis. Pepsin* (1 % by weight) is added to the protein in 5 % formic acid at pH 2 and hydrolysis is allowed to proceed at 40°C for at least 6 hr (overnight is often preferable). The peptide mixture is lyophilized and redissolved or suspended in 0.2 M NH,HCO, to give a final pH between 7.5 and 8.0 (adjust with NH3 if necessary). The pepsin is denatured by the high pH. Subtilisin* (1 % by weight) is added, and hydrolysis is allowed to proceed at 40°C for at least 6-8 hr (or overnight). The subtilisin is subsequently inhibited by incubation with 0.1 % (v/v) diisopropylfluorophosphate (DFP) for 1 hr at 40°C. In some instances it may be possible to omit the use of subtilisin after pepsin, or pronase or papain may be found to give better results. MnCl, (0.025 M) is added to the peptide mixture obtained by the above procedure (at pH 7.5 to 8.0) to a final concentration of 0.005 M and aminopeptidase M (2000 mU per ml of peptide mixture) and prolidase (imidodipeptidase; 1W200 units per ml of reaction mixture) are added; hydrolysis is allowed to proceed for at least 1 6 2 4 hr. The enzymic hydrolysate is lyophilized to dryness over P,O,, redissolved in 5 % acetic acid and lyophilized again. This procedure is usually sufficient to remove most of the NH3, so that analysis for the basic amino acids can be made. Repeated additions of water (or 5 % acetic acid) and lyophilization can be used to reduce the NH, content further, if desired. The times of hydrolysis and the amounts of enzymes added (2 or more additions of pepsin or subtilisin may sometimes prove beneficial)

* See Appendix 11.

Ch. 2


41

may be increased to give better hydrolysis if evidence is obtained for incomplete hydrolysis. Preliminary evidence for incomplete hydrolysis may be obtained by a peptide map of the hydrolysate (use at least 0.05 pmoles of the protein) using electrophoresis at pH 1.9 in the first direction (see Appendix) and chromatography in 1-butanol-acetic acid-water (200:30: 75, v/v) as the second direction. Comparison of this map with a map of a standard mixture of amino acids (plus any derivatives that may be present) may demonstrate additional ninhydrin-positive spots that may correspond to unhydrolyzed peptides. Unhydrolyzed peptides are sometimes also detected as additional peaks eluting from the columns of amino acid analyzers. Peptides frequently have poor color values and are not detected as easily as the amino acids. 2.11.2. Procedure 2 : use of insoluble enzymes

A newly described procedure for complete enzymic hydrolysis of proteins and peptides (Brown and Wold 1973) makes use of several proteases which are covalently attached to insoluble supports (Cuatrecasas et al. 1968).The advantages of using these enzymes in the insoluble form are that the enzymes may be easily removed from the products, and there is little or no hydrolysis of the enzymes themselves which reduces the corrections for controls. One disadvantage of the insoluble proteases may be that the size of the insoluble particles introduces steric factors hindering the hydrolysis of large peptides or proteins. Although the details of the hydrolysis procedure are yet to be published in detail (see Brown and Wold 1973),the first hydrolysis is with agarose-bound aspergillopeptidase A* at pH 1.5-2.0 for 18 hr. The resulting peptide mixture is then hydrolyzed simultaneously with agarose-bound pronase* and aminopeptidase M* for 16 hr in borate buffer at pH 7.6, and finally with agarose-bound prolidase* for 2 hr in the same buffer. It may prove helpful (not indicated in the procedure) to add the prolidase at the same time as the aminopeptidase M, since it

* See Appendix 11. Subject index p. 201

42


has been demonstrated that X-Pro dipeptides are often found in aminopeptidase M hydrolysates as diketopiperazines which are resistant to the action of prolidase (e.g. see Waley et al. 1970). The proteases are presumably removed by filtration or centrifugation (with washing). 2.11.3. Procedure 3 : papain, leucine aminopeptidase, prolidase This early procedure for enzymic hydrolysis of proteins was reported by Hill and Schmidt (1962) to be successful for hydrolysis of several proteins. Papain was found to be superior to subtilisin or a combination of trypsin and chymotrypsin for the initial hydrolysis. The method might be improved if aminopeptidase M (discovered after the method was developed) is used in place of the leucine aminopeptidase, but to our knowledge this has not been tested. The problem with diketopiperazine formation from X-Pro dipeptides in aminopeptidase M hydrolysates of peptides (see above) may make this substitution less desirable than it would seem at first. In this procedure, 1-2 pmoles of protein (0.1-0.5% in water) at pH 7.0 is heated at 95°C for 3 min to denature the protein. After cooling, buffer (0.2M sodium acetate at pH 5.2) is added to a final concentration of 0.03 to 0.05 M, and the mixture is incubated at 4OoC for 10 min. Sodium cyanide (0.1 M, neutralized to pH 7.0) is added as activator to a final concentration of 0.01 M, and papain (C, of 1.2 to 1.5)* in a 0.5 to 1.0% solution is added to give a final weight equal to 1 to 5 % that of the protein. After hydrolysis for 18-24 hr at 40°C in a closed container, the pH is adjusted to 2 to denature the papain, and a measured aliquot is lyophilized to dryness. The dried peptide mixture is redissolved in 1-2 ml H,O and adjusted to pH 8.4-8.6 with dilute NaOH. Tris buffer (other buffers can be substituted to prevent interference with any subsequent analyses) (0.5 M at pH 8.5) and MnCl, (0.025 M) are added to give final con-

* The specific activity (C,) is expressed as K , (first-order rate constant calculated in decimal logarithms) per ml of protein N per ml of reaction mixture and is determined at a substrate concentration of 0.05 M a-N-benzoyl-L-argininamide at pH 5.5 and 39°C in the presence of 0.005 M cysteine and 0.001 M EDTA.

Ch. 2


43

centrations of 0.005 M. Leucine aminopeptidase (4 to 6 mg, C, = 88) and prolidase (0.5 to 1.0 mg, C1=20) are added, and hydrolysis is allowed to proceed for 15 to 24 hr at 40°C. The enzymes are inactivated by adjusting the pH to 2 with 1 N HCl, and prior to analysis the amino acids are removed from the proteins by dialysis, gel filtration or other methods.

2.1 1.4. Analysis of enzymic hydrolysates Tryptophan, asparagine, glutamine and any other acid-labile derivatives will be the major components which are present in enzymic hydrolysates but not in acid hydrolysates. If cysteine and cystine are converted to S-cdrboxymethyl cysteine (9 3.8.2), this will be present in both the acid and enzymic hydrolysates. Tryptophan may be analyzed as indicated in 92.6. Asparagine and glutamine can be quantitated together as a single peak in many common analyzer systems employing a temperature change which separates the amides from the serine position where they elute in most routine systems. The color values for asparagine and glutamine are 93% and 87%, respectively, that of aspartic acid (Spackman et al. 1958). By the use of a lithium buffer system, asparagine and glutamine can be separated from each other and from other amino acids, thus allowing them to be quantitated individually (Benson et al. 1967). Acid-labile derivatives may elute from analyzer columns in the positions of other amino acids, and therefore, may be missed and interfere with accurate quantitation of the other amino acids. Modification of column temperature, buffer pH, buffer ionic strength, etc. may often be necessary to give good resolution of some of these derivatives.

2.12. Naturally-occurring amino acid derivatives in proteins An ever-increasing number of amino acid derivatives is found to occur naturally in proteins. These include E-N-methyllysines, E-N-acetyllysine, a-N-acetyl derivatives of various residues, N-methylhistidines, N-phosphohistidines, N-methylarginines, phosphoserine, phosphothreonine, hydroxyproline, hydroxylysine, trimethylhydroxylysine, tyrosine sulfate, iodotyrosines, carbohydrate moieties and coenzymes Subjecr inder p 201

44


covalently bound to various residues, lysine condensation products in proteins like elastin, adenyl groups covalently bound to tyrosine residues, covalently bound ADP-ribose moieties, etc. Many of these derivatives are acid-labile and would not be detected after acid hydrolysis of the proteins. To the investigator studying chemical modification of proteins, derivatives like those listed above are of interest for two primary reasons: (1) some derivatives are sites of modification and (2)complete modification of a particular class of amino acid functional groups with a reagent may not be possible because of the presence of these derivatives. For any protein being investigated by chemical modification and containing these types of derivatives, any results and conclusions must account for the effects of these derivatives. 2.12.1. Lysine derivatives 2.12.1. I . &-N-Acetyllysine This derivative of lysine has thus far been identified only in histones (see DeLange and Smith, 1971, 1974), but may also be produced in other proteins by chemical acetylation, e.g. with acetylimidazole, or acetic anhydride. Since the &-amidebond is acid-labile, this derivative may be present in other proteins, which have not been hydrolyzed and analyzed by enzymic methods (see tj 2.11). &-N-Acetyllysinemay be identified (see DeLange et al. 1969a) by paper electrophoresis at pH 1.9 (migration of 0.7 compared to 1.0 for alanine), paper chromatography in 1-butanol :glacialacetic acid :H 2 0 (200:30:75 v/v) (migration is 0.85 that of tyrosine) or 1-butanol: pyridine :glacial acetic acid :H 2 0 (90:60 : 18 :72 v/v) (migration is 0.89 that of valine). With the 60 cm column of a Beckman 120B analyzer, c-N-acetyllysine elutes at 77 min compared to 86 min for glycine and 92.5 min for alanine. We have used the average color value, calculated as indicated by the analyzer manufacturer, as the color value for estimating the amount of this derivative. (The average color value is obtained by averaging the color values for the amino acids resolved on the long column of the analyser with the exception of proline and cystine).

Ch. 2


45

2.12.1.2. c-N-Methyllysines The e-N-methyllysines have been found in histones (see DeLange and Smith, 1971, 1974), cytochromes c (see DeLange et al. 1970), flagellin (see Glazer et al. 1969), ribosomal proteins (Comb et al. 1966) and muscle proteins (see Paik and Kim, 1971 for a review). In some proteins (e.g. calf thymus histone 111) c-N-monomethyllysine, c-N-dimethyllysine and e-N-trimethyllysine are all three present, whereas in other proteins only one derivative (e.g. cytochromes c) or two of the derivatives (e.g. calf thymus histone IV) are present. Methylated lysines can also be formed in proteins by chemical modification in uitro (Q 3.1.1.3). The c-N-methyllysines are stable to acid hydrolysis and may be identified with the aid of an amino acid analyzer. With many routine analyzer systems E-N-monomethyl- and E-N-dimethyllysine elute as a shoulder or a partially resolved peak on the trailing edge of the lysine peak. If a 60 cm column is eluted with the buffer usually used for the 15 cm column these two methylated lysines are well resolved from lysine but not from each other. E-N-trimethyllysine usually elutes with lysine or slightly ahead of it in these systems (15 cm or 60 cm column eluted with pH 5.28 buffer). Paik and Kim (1967)developed a method utilizing the 60 cm column, eluted at 30°C with a citrate buffer (0.2 M) at pH 5.84, which resolved all three methyllysine derivatives from each other and from other basic amino acids. It has been found that the separations are sensitive to slight changes in pH, column temperature, ionic strength and the type of resin used. In our own laboratory we have used a 48 cm column of Durrum DC-2 resin (Appendix 11) (DeLange et al. 1970) which gives excellent resolution, but requires at least 7 hr (1 8 hr if arginine analyses are required). Seely et al. (1969) have used a column (0.9 x 15 cm) of Aminex A-5 resin (Appendix 11), eluted at 25°C with buffer at pH 6.48 (prepared by adding NaOH to the usual pH 5.28 buffer) to give resolutions in 2 hr (or even less) that should be adequate for many purposes. In this system at a flow rate of 34 ml per hr. the elution volumes and color constants (in parentheses) were :lysine, 47.5 ml(43.5); c-N-monomethyllysine, 55 ml (39.7); E-N-dimethyllysine, 59.5 ml (39.2); and Subject index p. 201

46


e-N-trimethyllysine, 63 ml(35.5). Histidine elutes just before lysine and NH3 elutes just after e-N-trimethyllysine. All of the acidic and neutral amino acids elute before 28 ml. Benoiton et al. (1971) have found that lysine and E-N-monomethyllysine have similar 570/440 nm ratios after reaction with ninhydrin as measured with the analyzer, and that 8-N-dimethyllysine and E-Ntrimethyllysine also have similar 570/440 nm ratios, but the 570/440 nm ratios of the two pairs are different. This provided a method for estimating from the 570/440 nm ratio the content of e-N-monomethyllysine and e-N-dimethyllysine in hydrolysates that are analyzed in systems in which these two derivatives are not separated. For other analyzer systems that have been used to study methylated lysines, see the references cited above and also Deibler and Martenson (1973). The methylated lysines can also be separated and identified by paper chromatographic systems. However, in electrophoresis at pH 1.9 all three derivatives migrate similarly to arginine. They also migrate with lysine during chromatography in 1-butanol :glacial acetic acid:water (200;30:75 v/v), but are partially resolved in 1butanol : pyridine :glacial acetic acid :water (90:60 :18 :72 v/v) for 40 hr :lysine and E-N-trimethyllysine, 1.OO;e-N-monomethyllysine and E-N-dimethyllysine, 1.18;and arginine, 1.38(see DeLange et al. 1969b). The best paper resolutions of the methyllysines are obtained by descending chromatography for 16-18 hr at room temperature (21"C), using rn-cresol: 88% phenol :borate buffer (190: 165 :45 v/v) with papers dipped (and then dried) in 0.1 M EDTA (adjusted to pH 7.0 with NaOH) essentially as described by Stewart (1963) except that Whatman No. 3 paper (57 cm in length) can be used in place of No. 52 paper, and the papers do not need to be dipped in acetone. The borate buffer is prepared by dissolving 3.944 g of boric acid in water, adjusting the pH to 9.3 with NaOH and diluting to 1 liter. Because of the pH 9.3 buffer used, better colors are obtained with a ninhydrin reagent buffered at a lower pH (e.g. 0.1% ninhydrin in 95 % ethanol : glacial acetic acid : collidine (50 : 15 :2)). The R, values are iysine, 0.18 ; arginine, 0.30 ; histidine, 0.43 ; e-N-monomethyllysine, 0.44; E-Ndimethyllysine, 0.76; E-N-trimethyllysine,0.88 (DeLange et al. 1969b).

Ch. 2


47

Histidine and E-N-monomethyllysine can be distinguished by their different colors with ninhydrin (supplemented by the use of Pauly reagent) or by using two-dimensional separations (e.g. pH 1.9 electrophoresis, followed by the cresol-phenol system). 6-N-Monomethyllysine has a color with the ninhydrin reagent similar to that of lysine (blue-grey) whereas the other two derivatives give colors similar to arginine (blue-violet), and histidine gives a yellow-brown color. 2.12.1.3. Hydroxylysine

Hydroxylysine is found in collagen and gelatin and possibly some other other proteins (see Vickery 1972). It is formed in uiuo by oxidation of lysyl residues in these proteins at the b-carbon. Hydroxylysine is stable to acid hydrolysis, but undergoes racemization under the usual conditions of hydrolysis to give a mixture of the natural hydroxylysine and allo-hydroxylysine. Several analytical systems appear to be satisfactory for the resolution and quantitation of hydroxylysine. If a protein is suspected of having hydroxylysine, perhaps several systems should be used. Spackman et al. (1958)resolved the hydroxylysines from a number of other components on a 50 cm column of Amberlite IR-120, eluted at 30°C with 0.38 N sodium citrate buffer at pH 4.26 ; hydroxylysine, 131 ml ;allo-hydroxylysine, 137 ml; and lysine 218 ml. The hydroxylysines elute after the amino sugars in this system. Knox et al. (1970) showed that hydroxylysine is eluted just before lysine from a 0.9 x 7 cm column of PA-35 resin, eluted with 0.255 M sodium citrate buffer at pH 5.0 (see also Hugli and Moore 1972).An early system developed by Piez and Morris (1960) was later modified (Miller and Piez 1966)to give an accelerated method for analyzing for all of the constituents of collagen. The resins (PA-28 at 55°C or AA-15 at 60°C) were eluted in a column (0.9 x 51 cm) at 80 ml/hr with a gradient established with a Varigrad containing various amounts of pH 2.91 citrate buffer and 0.4 M sodium citrate in the chambers. Hydroxylysine elutes at about 213 ml, between phenylalanine (197 ml) and ammonia (246 ml). The color value for hydroxylysine is about the same as that of lysine (Spackman et al. Subject index p . 201

48

CHEMICAL MODIFICATION Db PROTEINS

1958). Hydroxylysine has a mobility of 2.26 compared to aspartic acid in p H 1.9 electrophoresis. 2.12.1.4. Trimethylhydroxylysine ( T H L ) and its phosphate ( T H L P ) These two derivatives have been identified in proteinacious material from cell walls of diatoms (Nakajima and Volcani 1970). In pH 1.9 electrophoresis they migrate 2.4 cm (THLP) and 6.1 cm (THL) towards the negative electrode,respectively,at 33.3 volts/cm for 20 min. In the analyzer system of Dus et al. (1966), THLP elutes in about the same position as threonine, and THL elutes between NH, and arginine (a buffer peak also elutes in this region). In the complex analyzer system of Hamilton (1963), THLP elutes at a volume about 2.4 times that of cysteic acid and 0.58 that of aspartic acid. 2.1 2.1.5. Desmosine, isodesmosine and related derivatives Lysine (and hydroxylysine) residues in collagen and elastin are -NH

CO-

\ /

FH

(YH2)3 CHO ~-1ysylal

+

-co‘CH-(CH,)2-(?-l:,

/

CHO

‘co-

/

-NH

e-lysylal

CHO

NH-

FH2-(CH,),-CH

y H2

e-lysylal

(YH2)4

-N
co

Two isomeric c-N-acyl derivatives of lysine are produced, differing in pK,. These have a half-life of approximately 20 min at pH 2 and 20°C The E-N-acyl derivative of lysine has a half-life of less than 3 min at pH 3.5 and 20°C Two isomeric c-N-acyl derivatives are formed at the same rate. These have a half-life of 4-5 hr at pH 3 and 25°C

aco\o / co

'

F2C-CO, F*C-CO

/o

Half-lives of derivatives are the same as for those obtained with the tetrahydrophthalic anhydride. This reagent does not alkylate sulfhydryl groups

E-N-acyl derivative is rapidly hydrolyzed at pH 9.5 and 0°C

* Both CI- and c-NH, groups are modified. Acyl derivatives of tyrosine formed by these reagents hydrolyze spontaneously under mild alkaline conditions. Small amounts of stable acyl derivatives of serine and threonine are formed. These can be quantitatively decomposed by treatment with 1 M NH,OH at pH 8.2 at 25°C for 2 hr. The olefinic reagents alkylate sulfhydryl groups with the formation of stable derivatives. ** Dixon. H. B. F. and R. N. Perham (1968) Biochem. J . 109, 312. ' Riley. M. and R. N . Perham (1970) Biochem. J. 118,733. t t Braunitzer. G.. K. Beyreuther. H. Fujiki and B. Schrank (1968) 2. Physiol. Chem. 344. 265

Subject mde-r p 201

84


produced by the arrangement of the carboxyl groups adjacent to each other on the hydrophthalic ring.

3.2. Modifications of arginine Several procedures are available for the modification of arginine residues. A partial list of reagents for derivatization of arginine would include butanedione (Yankeelov et al. 1968),phenylglyoxal (Takahashi 1968), malondialdehyde (King 1966) and nitromalondialdehyde (Signor et al. 1971). None of the procedures are ideal since the first two reagents yield products which at present are incompletely characterized while the conditions of derivatization with malondialdehyde and nitromalondialdehyde require strongly alkaline conditions. Two important motivations exist for the modification of arginine residues. The first would be to restrict the sites of proteolytic digestion by trypsin to lysine residues. The second would be to determine the potential functional importance of an arginine residue in the mode of action of a given protein. The best reagents for the latter purpose are 2,3-butanedione and phenylglyoxal. 3.2.1. Butanedione reaction with arginine Yankeelov et al. (1968)have reported that the trimer of2,3-butanedione (I) as well as the dimer (11) are the reactive forms of this reagent. Reaction conditions for the exhaustive modification

I

11

of protein arginine with trimeric 2,3-butanedione involve incubation of the 0.4 M reagent I and protein at 25OC in 0.5 M phosphate buffer

Ch. 3

MODIFICATION OF PROTEIN SIDE-CHAINS

85

(pH 7.0) for 48 hr. The synthesis and structure of the derivative formed have not been reported but whatever its nature, it does not regenerate arginine upon acid hydrolysis. Since the reagent readily reacts with amino groups at the relatively high concentrations required for complete derivatization of the protein, protection of the &-aminogroup is essential if specific cleavage by trypsin at lysine residues is desired. A possible method of protection prior to derivatization with compound I would involve citraconylation of the lysine residues prior to reaction. Riordan (1973) has reported that the monomer of 2,3-butanedione inactivates carboxypeptidase as effectively as the trimer. One interesting feature of his study is that a 0.05 M borate buffer enhances the rate and possibly affects the distribution of products formed from both monomeric and trimeric butanedione when compared with a 0.05 M veronal buffer at the sample pH. He has attributed this specific buffer effect to the formation of a cyclic borate ester following the initial condensation of the guanidinium group with 2,3-butanedione as indicated in eq. (3.1). The conditions of modification with butanedione used by Riordan (1973) involved incubation of the protein at pH 7.5 at 20°C for 15-60 min at concentrations ofbutanedione ranging from 2.2 x lo-’ M to 7.5 x lo-’ M. CH3

R I

\

c=o I

NHO

NH

+

I I1

2\

C--N-(CH~)~-C-C-O@

HN//

A

H

CH3

Subject index p. 201

86


3.2.2. Phenylglyoxal reaction with arginine Phenylglyoxal was first introduced as a protein modification reagent by Takahashi (1968). The product obtained with either arginine or N-acetyl arginine contains two moles of phenyl glyoxal per mole of amino acid. The precise structure of the derivative formed has not been determined but two suggested structures with N-acetyl arginine are indicated below. Whatever the structure of the product, arginine is not regenerated by acid hydrolysis. Further, the product is unstable even at slightly alkaline pHs. For example, 66 7; of the phenylglyoxal derivative of arginine is regenerated after 48 hr upon incubation at 25°C in 0.1 M N-ethylmorpholine acetate buffer, pH 8.0.

?

OH

OH

I

N

,N-CH-C

I

7

II

OH 0

H-C-N-R

I

C

\o-

I

NH

I

(CH2)3

I

HC-NHR I

c-oII

0

As a general procedure for the modification of proteins by phenylglyoxal, Takahashi (1968) suggested reacting a 0.3 to 3 % solution of phenylglyoxal with the protein in 0.2 M N-ethylmorpholine acetate buffer pH 8.0 a t 25°C. Under these conditions, N-a-dinitrophenyl arginine completely reacts in 20 hr. An important side reaction which limits the usefulness of phenylglyoxal as a modification reagent is its ability to deaminate free amino acids and the N-terminal amino

Ch. 3


87

group of peptides. Further, the lability of the phenylglyoxal derivative of arginine in alkaline solution requires mildly acidic conditions to be employed if the isolation of peptide containing the modified arginine residue is desired. 3.2.3. Nitromalondialdehyde reaction with arginine The products formed by nitromalondialdehyde have been characterized (Signor et al. 1971). Thus N-acetylarginine reacts with nitromalondialdehyde to yield a nitropyrimidine derivative as indicated below in eq. (3.2.).

? I ? C-C-ONH

CHO

HN,

+

02N-& ‘CHO

/ C-N-CHz-CH2-CH H2N

2-H

b

Since the optimum pH range for the reaction is between pH 12 and 14, the usefulness of nitromalondialdehyde is restricted to denatured proteins. To modify the S-carboxymethylated B chain of insulin, Signor et al. (1971) added about 3.0 mmoles of the peptide dissolved in 1.0 ml of 1 M NaOH to 1 ml of water containing 10 mg of nitromalondialdehyde (70 mmoles). After 2 hr at 0-5°C excess modification reagent was removed by gel filtration using 5 0 x formic acid as eluent. Arginine residues modified by nitromalondialdehyde are resistant to tryptic digestion. However, reduction of the derivative by sodium borohydride yields tetrahydropyrimidyl compounds which are susceptible to hydrolysis by trypsin. Modification by nitromalondialdehyde thereby permits the restriction of tryptic digestion to lysine Sah,t,c/ rnder p 201

88


residues but yet would allow the further tryptic cleavage of the resulting peptides after reduction with sodium borohydride.

3.3. Modifications of carboxyl groups The two reactions most commonly employed for the modification of protein carboxyl groups have been acid-catalyzed esterification (Wilcox 1967), and coupling with nucleophiles mediated by a watersoluble carbodiimide. The latter reaction has been exploited widely in the determination of the carboxyl group content of proteins as well as in studies of carboxyl group function, and would at present appear to be the reaction of choice for these purposes. The following method has been described by Carraway and Koshland (1972) for the determination of carboxyl group content of proteins. The protein (10 mg/ml) and glycinamide HCl (1 M) are dissolved in guanidine hydrochloride (6 M). The pH is adjusted to 4.75 with 1 M HC1 using a pH stat, and sufficient solid ethyldimethylaminopropylcarbodiimide is added as a solid to bring its concentration to 0.1 M. The pH is maintained by the automatic addition of M HCl. The reaction is terminated after 1 hr by the addition of an excess of 1 M sodium acetate buffer at pH 4.75 which reacts with excess carbodiimide. The residual reagents are removed by exhaustive dialysis. All traces of free glycinamide must be removed. The modified protein is then subjected to amino acid analysis. The difference in glycine content of the protein before and after modification represents the carboxyl group content. This procedure offers considerable flexibility in the choice of nucleophile. For instance, norleucine, or taurine, or 14C-labelled glycinamide or glycine methyl ester may be employed in the place of glycinamide. It should be noted that the carbodiimide reacts with both sulfhydryl and tyrosyl OH groups as well as with carboxyl groups. Tyrosine can be regenerated by treatment of the derivative with 0.5 M hydroxylamine at pH 7 and 25°C for 5 hr. Successful regeneration of thiol groups has not been reported. The reduction of carboxyl groups and their reaction with diazonium

Ch. 3

89

MODlFlCATlON OF PROTEIN SIDE-CHAINS

salts, isoxazolium salts, triethyloxonium fluoroborate, have all been exploited to a limited extent. These reactions are discussed in the reviews listed in ch. 1.

3.4. Modifications of histidine Reagents capable of modifying the histidyl residue with complete specificity are not available to date. Reaction with a-haloacids and amides at near neutral pH offers the best approach to the modification of histidine in native proteins. In a protein such as insulin, which contains neither methionyl nor cysteinyl residues, reaction with iodoacetate at pH 5.6 leads to the formation of a derivative in which the sole modification is the N-carboxymethylation of two histidyl residues (Covelli et al. 1973). Alkylation of histidyl residues leads to derivatives substituted on either the N-1 or N-3 imidazole nitrogen, as well as to the disubstituted derivative (Crestfield et al. 1963; Heinrikson et al. 1965). y

TH2

2

C==rCH2-CH-coo N

N ~ C H 2 - C H - C o o H V N‘CH~COOH

HOOCCH,

/N

3-carboxyrnethyl histidine

1-carboxyrnethylhistidine

N

HOOCCH~’

W

N< CH,COOH

1, 3-dicarboxyrnethylhistidine

When an unusually reactive histidyl residue, for example at the active site of an enzyme, is alkylated, the reaction tends to be strictly stereospecific, as observed in ribonuclease (Crestfield et al. 1963) or Subjrcr inder p . 201

90


carbonic anhydrase (Bradbury 1969), and only one of the above derivatives is generally formed. However, alkylation of histidyl residues in denatured proteins gives rise to all three products, as does alkylation of native proteins with large excess of reagent (Harris and Hill 1969). In such cases, any histidyl residue may give rise to unequal amounts of derivatives substituted on either of the imidazole nitrogens. The alkylation of methionine is pH-independent, while that of histidyl, cysteinyl and lysyl residues is pH-dependent, the rate reflecting the concentration of the unprotonated protein nucleophiie. Dye-sensitized photooxidation (Ray 1967 ;Westhead 1972) has been employed successfully in a number of cases (e.g. Forman et al. 1973) to obtain selective destruction of histidyl residues. Other reagents which have been used for the modification of histidyl residues in proteins include 1-fluoro-2,4-dinitrobenzene, diazonium1H-tetrazole (see Andres and Atassi 1973), iodine (Covelli and Wolff 1966 ; Wolff and Covelli 1969), N-bromosuccinimide, ethoxyformic anhydride (Melchior and Fahrney 1970), and bromoacetone (Beeley and Neurath 1968).These reagents appear to offer little advantage over alkylation with a-haloacids, or dye-sensitized photo-oxidation. Indeed, some are far less specific.

3.5. Modifications of methionine Reagents reactive solely toward the methionyl side-chain in natiue proteins have not been described to date. Selective conversion of the thioether to the sulfoxide derivative by photosensitized oxidation has been reported in a protein devoid of free thiol groups. The selective photosensitized oxidation employed methylene blue or hematoporphyrin as sensitizers, and aqueous acetic acid (3Q-90;, v/v), or acidic buffers at pH 2-6.5, as solvents (Scoffone et al. 1970). Anaerobic photo-oxidation in 4 M aqueous acetone has been reported to lead to the same result (Gennari and Jori 1970).Methionine can be regenerated from the sulfoxide by incubation with thiols ( 5 '/, aqueous b-mercaptoethanol at pH 8.0, for 24 hr under N,; Jori et al. 1968). The reaction of iodoacetic acid and iodoacetamide with proteins at

Ch. 3


91

pH < 4 is limited to cysteinyl and methionyl residues. Methionine is converted to methionyl sulfonium salts which give a number of products upon acid hydrolysis, thus complicating the quantitation of the extent of derivative formation (55.3.2.2). The most direct way of estimating methionine sulfonium salt is by performic acid oxidation (ij 3.2.5.1 and 4 3.8.1). Methionyl residues are converted to the sulfone, while methionyl sulfonium salts give rise to several decomposition products which do not include methionine sulfone (Goren et al. 1968; Naider and Bohak 1972; $5.3.2.2). Naider and Bohak (1972) have found that incubation of sulfonium salts with thiols leads to the regeneration of methionine. Complete recovery of methionine was obtained upon incubation of the Scarboxyamidomethylmethionyl sulfonium salt with O. 12 M mercaptoethanol at pH 8.9 for 24 hr at 30°C. In contrast, the S-carboxymethylmethionyl sulfonium salt is regenerated to only a minor extent under these conditions. Parenthetically, it may be noted that the S-carboxyamidomethylmethionyl residue may be used as a site of cleavage of the polypeptide chain (in a reaction analogous to that obtained with CNBr) by boiling in water for 2 hr (Tang and Hartley 1967, 1970).

3.6. Modifications of tryptophan The reagents employed for the selective modification of tryptophan all react with cysteinyl residues. Hence, either modification or protection of these residues is necessary, and is ussumed in the subsequent discussion.

Recently, 2-(2-nitrophenylsulfenyl)-3-methyl-3-bromoindolenine (BNPS-skatole) has replaced N-bromosuccinimide, which had been used frequently in earlier studies. At low reagent to protein tryptophan ratios, in 50 ”/, aqueous acetic acid, BNPS-skatole reacts selectively with tryptophan residues converting these to the oxindole derivative. Methionine is concommitantly converted to the sulfoxide. At high concentrations of reagent, slow selective cleavage (to the extent of 15-60%) of the peptide bonds involving tryptophanyl residues is Sublecr rnderp 201

92


obtained. The experimental details concerning the preparation and use of this reagent are discussed by Fontana (1972). Various nitrobenzylhalides have been widely used for the modification of tryptophan as well a5 for the determination of the content of this amino acid in proteins (Horton and Koshland, 1972). The reaction of these compounds with tryptophan and its derivatives gives rise to complex mixtures of products (Horton and Koshland 1972; Heinrich et al. 1973). Other procedures which have been applied to the modification of tryptophan include ozonolysis in anhydrous formic acid containing resorcinol (Previero et al. 1967a)giving rise to the N-formylkynurenine derivative, and proflavine-sensitized photooxidation in 98-100 acid (Scoffone et al. 1970). Two apparently highly selective procedures have been developed in recent years : formylation and sulfenylation. 3.6.1. Formylation of tryptophan

The procedure outlined below, based upon the description of Previero et al. (1967b)has been used successfully,with minor variations for the selective modification of tryptophan residues in lysozyme (Previero et al. 1967b), trypsin (Coletti-Previero et al. 1969), cytochrome c (Aviram and Schejter 1971), and thioredoxin (Holmgren 1972). Protein (10-20 mg) is dissolved in 1 ml of formic acid (98-100%) and 1 ml of formic acid saturated with HCl is added. The mixture is incubated in a glass-stoppered tube at room temperature. At 20 min intervals, 50 pl aliquots of the solution are pipetted into 2.5 ml of 8 M urea-0.05 M sodium acetate, pH 4.0, and the absorbance at 298 nm determined. ( E f~o r 1-formyltryptophan in 8 M urea at pH 4 is 4880.) Hence, the degree of modification can be calculated from the total

Ch. 3

93


change in absorbance at 298 nm and the known concentration of protein). When no further change in absorbance at 298 nm is observed, the reaction mixture is diluted 5-fold with ice-cold water, and dialyzed against water at 4°C. The modified protein is then recovered by lyophilization. 1-Formyltryptophan is stable up to pH 7 at room temperature. The decomposition of the derivative back to tryptophan at alkaline pH forms the basis of the deformylation procedure. The protein derivative (approx. 5 mg/ml) is dissolved in 0.05 M NH,OH-NH,Cl, pH 9.5, containing 6 M guanidine hydrochloride. After incubation at room temperature for 16 hr, the protein is dialyzed and lyophilized. 3.6.2. Sulfenylation of tryptophan The reaction of arylsulfenyl halides with proteins at p H d 3.5 is limited to the cysteinyl and tryptophanyl residues (Fontana and Scoffone 1972), as illustrated below for 2-nitrophenylsulfenyl chloride (NPS-Cl). I CH2SH -NH-CH-CO-

+

6I

CH2-CH

H

N0 2

,NH-

CH2-S-S -NH-CH-COI

-Q NO2

sc 1

/NH-

H NO2

In proteins devoid of - SH groups, or those in which these groups have been alkylated or otherwise modified (5 3.8), the reaction is restricted to tryptophanyl side-chains. The protein (10 mg/ml) is dissolved in either water or 5 % acetic acid. NBS-Cl (4 moles per mole of tryptophan) dissolved in glacial acetic acid is added, with vigorous stirring, to give a final acetic acid concentration of 25 %. After 6 hr in the dark at room temperature the reaction Sublecl mde.rx p . 201

94


is terminated by passage of the mixture through a column of suitable size of Sephadex G-25 equilibrated with 30 acetic acid. The proteincontaining fractions are pooled. Measurement of the absorbance at 365 nm permits the calculation of the concentration of the NPS~ ~ L. mole~ cm- ; Scoffone tryptophan derivative present ( E = 4000 et al. 1968), while the protein concentration is determined on an appropriate aliquot by amino acid analysis. The extent of sulfenylation is then calculated from the relative concentrations of the protein and the NPS-derivative in the fraction. Site-specific sulfenylation of tryptophan residues in egg-white lysozyme has been attained by either limiting the amount of reagent, or modifying the reaction conditions. Thus, Trp-108 is the major site of modification upon addition of one equivalent of 2-thio-(2-nitro-4carboxypheny1)sulfenyl chloride to lysozyme in 25 acetic acid (Veronese et al. 1972). Specific modification of Trp-62 was attained in approximately 8 hr by the addition of 5 consecutive portions of solid NPS-Cl (10 pmole for each ml of reaction mixture) to a solution of lysozyme (0.5 pmole/ml) in 0.1 M sodium acetate at pH 3.5 (Shechter et al. 1972). In both cases, the protein derivative was separated from other products, and unreacted enzyme by ion-exchange chromatography. It may be noted, parenthetically, that a novel reaction of NPS-Cl with the thioether bonds linking the heme to cytochrome c leads to quantitative release of the heme (Fontana et al. 1973). -NH-CH-CO-

I

YHZ

S

I CH3-CH

NPS-C 1 66% acetic acid

-NH-CH-COI CH A

I S-S-CBH~ NO:,

Ch. 3

95


3.7. A4odiJi:cationsof tyrosine Modification of tyrosyl groups in proteins has been performed with acetylimidazole (Riordan and Vallee 1972a), N-bromosuccinimide (Spande et al. 1970), cyanuric fluoride (Gorbunoff 1972), diazonium compounds, such as diazonium-1H-tetrazole (Riordan and Vallee 1972b), tetranitromethane (Riordan and Vallee 1972c),and iodine (or other iodinating reagents). In each case, the specificity of these reagents towards tyrosyl side-chains is far from complete. In native proteins, selective modification may be achieved with any one of these reagents as a consequence of the microenvironment of a given residue. Clearly, this is not predictable in advance. The two most widely utilized procedures for tyrosine modification, nitration and iodination, are discussed below. 3.7.1. Nitration of tyrosine 0-

0-

I

Nitration of tyrosyl residues with tetranitromethane (TNM) is now one of the most frequently attempted modification reactions for native proteins (see Riordan and Sokolovsky 1971fora review and references). The mechanism of this reaction has been examined by Bruice et al. (1968),who propose the following scheme : XCgH40-

t T N M e

charge-transfer

-rate-determining step

02NXCsH40-

t

IXC,H,O. XCsH40

t NOZ,l

4 t NO,, '

+ C(NOZ)J nitro-formate anion

NO2' t X C s H 4 0 - e X C s H 4 0 ' t NO, nXC6H40 ' -coupling

products

Awareness of this reaction sequence is of importance to the interpretation of the results of the reaction of TNM with proteins. Subject mderp. 201

96


The potential advantages of selective nitration of tyrosyl residues in native proteins are numerous. The reaction is performed under mild conditions, giving rise to a 3-nitrotyrosyl derivative (pK’- 7), which in the acid form absorbs intensely at 3-50 nm. Hence, the nitrotyrosine content may be readily determined spectrophotometrically, as well as by amino acid analysis (8 2.2.3). The absorption spectrum of 3-nitrotyrosine is highly sensitive to solvent polarity and exhibits significant optical activity in the long wavelength absorption band. Consequently, nitrotyrosyl residues can be utilized as indicators of conformational change, or of interactions of proteins with other macromolecules or small molecules (e.g. Kirschner and Schachman 1973). Any perturbation in the pK, of nitrotyrosyl residues is readily determined spectrophotometrically. 3-Nitrotyrosine is readily reduced to 3-aminotyrosine by low concentrations of sodium hydrosulfite (Na,S,O,). The pK, of the phenolic group in the latter derivative (pK’= 10.1) is similar to that of tyrosine - OH, while the pK, of the 3-amino group is 4.7. Reduction permits, therefore, a distinction to be made between the effects of the initial nitration reaction, due to the perturbation of the pK, of the phenolic - OH, and those due to steric hindrance or conformational change resulting from the presence of a substituent on the phenolic ring. The 3-aminotyrosyl residue offers opportunities for selective reaction based upon the low pK, of the aromatic amino group (e.g. Cuatrecasas et al. 1969; Kenner and Neurath 1971), and metal ion chelation. The awareness of the many advantages of this modification reaction should be matched by the knowledge of the numerous side-reactions observed to result from the reaction of TNM with proteins. In addition to nitration of tyrosine, the following side-reactions have been reported in several (but not all) proteins studied : (1) inter- and intramolecular cross-linking, (2)oxidation of sulfhydryl groups to a variety of products, (3) oxidation of methionine, (4) modification of tryptophan and histidine, (5) modification of prosthetic groups. It is apparent therefore that a successful application of TNM to the selective modification of tyrosine is achieved in those cases where an unusually rapid reaction

Ch. 3

97


occurs with specific residues, owing to some factor such as complex formation with the reagent, or, where much effort and care is given to the purification and characterization of a homogeneous component from a complex mixture of products (e.g. Cheng and Pierce 1972). As expected from the reaction scheme shown above, the rate of nitration is strongly pH dependent. Some representative results obtained with carbonic anhydrase are shown in Fig. 3.1. It should be

I

I

I

50

I00

I50

.-c e

3.0 L

a r 0

e a

0

-e 0

. E

2.0

a e

c ._ In

L 0

c ) . 0 + L

-c

I .o

0

VI 0 3

-

Tt

Y)

0

a

0

Time o f r e a c t i o n ( m i n . )

Fig. 3.1. Reaction of human carbonic anhydrase B with tetranitromethane. The reaction mixture contained 5 ml of 0.3% protein and 25 p1 of tetranitromethane, buffered at pH 7.1 with 0.1 M sodium phosphate, or at pH 8.4 and 10.2 with 0.1 M Tris-H,SO,. (Data from Dorner 1971). Suhlect tndrPr p 201

98


noted that conformational change in response to alteration in pH is commonplace in native proteins and that the rate and extent of a chemical modification reaction may well be strongly influenced by this factor. A procedure applicable to the nitration of a native protein is outlined below. The protein (2-4 mg/ml) is dissolved in 0.1 M potassium phosphate buffer, pH 7.5, containing M EDTA.A 0.15 M solution ofTNM is prepared in 95 3' ; ethanol, and 20 p1 is added per ml of protein solution (final TNM concentration 0.003 M).After 60 min at room temperature the reaction is terminated by the addition of 50 pl of 1 M mercaptoethanol, followed immediately by exhaustive dialysis against 17; NH,HCO, at 4°C. The dialysis removes excess TNM and the colored nitroformate anion, which is a product both of the nitration reaction and of decomposition of the reagent. The dialyzed protein derivative is lyophilized. The visible absorption spectrum of a solution containing a known concentration of nitrated protein is measured in a solution buffered at pH 9.0, and the absorbance at the maximum (near 428 nm) used to calculate the nitrotyrosine content ( E for the ~ nitrophenoxide ~ ~ ~ion is 4200).The tyrosine and nitrotyrosine content of the modified protein should also be determined by amino acid analysis. If the sum of these values does not add up to the tyrosine content of the unmodified protein, intra- or intermolecular cross-linking may have occurred. The amino acid analysis may also reveal whether other side-reactions have taken place. Particular attention should be paid to the half-cystine, cysteine, methionine, histidine and tryptophan contents of the modified proteins. Polyacrylamide gel electrophoresis in the presence of sodium dodecyl sulfate offers a rapid and highly sensitive way of detecting products of intermolecular cross-linking. Such products are readily removed by gel filtration. It should be recognized that variations in pH, TNM concentration, or reaction time, from those given above, may be required to produce the desired degree of substitution in a particular case. Some recent interesting applications of nitration may be found in the

~

Ch. 3


99

work of Piszkiewicz et al. (1971),Dorner (1971), Hugli and Stein (1971), and Cheng and Pierce (1 972). 3.7.2. Iodination of tyrosine The iodination of proteins has long been a very popular modification reaction. This reaction, performed under mild conditions, has been applied to such varied ends as the detection of side-chains at the catalytic sites of proteins, investigation of the relative reactivity of tyrosyl side-chains, and the introduction of heavy atoms in connection with X-ray diffraction studies. The availability of two radioactive iodine isotopes, I z 5 I with a half-life of 56 days, and 13'1 with a half-life of 8 days, has been very extensively exploited in the preparation of hormone derivatives of high specific activity for radioimmunoassay and receptor studies. The reaction of proteins with HOI, generated either from triiodide, iodine monochloride, iodine plus chloramine T, etc., may involve a number of side-chains. In native proteins, halogenation of tyrosine takes preferentially, with the formation of the 3-iodo- and 3,5-diiodotyrosyl derivatives - histidine is halogenated less readily to give the 1-iodo- and 3-iodohistidyl derivatives as well as diiodohistidine. In certain cases (see Glazer 1970),reaction with thiol groups leads to the formation of 'stable' sulfenyl iodides (RSI). In addition to these substitution reactions, oxidative reactions involving methionyl, cysteinyl, and tryptophanyl side-chains may also take place. The substitution reactions predominate at alkaline pH, the oxidative ones at acid pH. For procedures for non-enzymatic iodination of proteins, most frequently carried out with iodine monochloride, reference should be made to the lucid treatment by Roholt and Pressman (1972). Recently, application of lactoperoxidase-catalyzed iodination has taken precedence over non-enzymatic procedures. The enzymecatalyzed reaction is reported to occur through the formation of an enzyme-substrate complex between the protein substrate and lactoperoxidase. No detectable free I, is formed by lactoperoxidase, and selectivity in attaining iodination of surface tyrosine residues in Sublecr index p 201

100


proteins without oxidative side-reactions is believed to be high. The method described by Morrison and Bayse (1970) for the enzymic iodination of tyrosine can be readily adapted to the modification of proteins. The reaction mixture contains, in order of addition, L-tyrosine (8.1 x M), KI (1.0 x M), lactoperoxidase (7.4 x M), in 0.05 M K-phosphate buffer, containing 1 x l o w 3M EDTA,at pH 7.4. The iodination is initiated by the addition of H,O, to a concentration of 1.0 x l o p 4 M. The specific activity observed for lactoperoxidase under these conditions was 1.05 x lo4 moles of L-3iodotyrosine per min per mole of enzyme at 25OC. At pH 7.4, the rate of enzymatic conversion of L-3-iodotyrosine to L-3,5-diiodotyrosine was 0.34 that of monosubstitution (Morrison and Bayse 1970).The desired level of iodination can be attained by successive equimolar additions of KI and H 2 0 2 to the reaction mixture. In this manner, only a low concentration of H, 0, is maintained, minimizing oxidation reactions. The concentration of lactoperoxidase may be calculated from the millimolar extinction coefficient of 114 at 412 nm, while the concentrations of stock H 2 0 2solutions may be determined from the absorbance at 230 nm and a molar extinction coefficient of 72.4 (Phillips and Morrison 1970). When radioactive iodine is to be used, it should be purchased as Na'25 I, free of carrier, or reducing agents. Picogram quantities of radio-iodine labelled proteins can be detected readily. Hormones, iodinated by the lactoperoxidase procedure to high specific activity, show retention of biological activity superior to that observed for preparations labelled by non-enzymatic methods (e.g. Frantz and Turkington 1972). A micro-scale iodination procedure has been described by Thorell and Johansson (1971). The reaction is performed in an 11 x 55 mm polystyrene tube at 22"C, and the reactants mixed continuously with a magnetic stirrer. The reagents and polypeptides are dissolved in 0.05 M sodium phosphate buffer, pH 7.5. NalZ5I (carrier-free, without reducing agent) is diluted with the buffer to an activity of -0.1 mCi/pl. The reactants are added in rapid succession in the following order and amounts: 0.5-1.8 mC Na 125 I(8-15 pl), 5 pg (25 pl) of the polypeptide

Ch. 3

101


or protein to be labelled, 4 pg (1.5 pl) lactoperoxidase, 1 p1 0.88 mM H 2 0 2 .The reactants are mixed for 30 sec, after which the contents of the tube are diluted with 500 pl of 0.05 M sodium phosphate buffer, pH 7.5, and immediately transferred to a column of Sephadex G-50 Fine (1 x 15 cm), equilibrated with 0.075 M sodium barbital buffer, pH 7.5. (The column is presaturated by the passage of 1 ml of a 1 % solution of bovine serum albumin). The first peak of radioactivity is collected and after addition of 0.5 ml of 1 % bovine serum albumin per ml it is stored at -20°C. For the determination of specific activity, an aliquot of the reaction mixture is transferred to a small dialysis tubing and dialyzed for 1 hr at 4°C against two changes of 1000 mlO.05 M phosphate buffer, pH 7.5, containing 5 g of Amberlite Resin IRA 401. The radioactivity of the reaction mixture is measured before and after dialysis. For a number of proteins, over 80% of the iodine added appears to be covalently bound to the protein. A water-insoluble lactoperoxidase, linked to CNBr-activated Sepharose beads, capable of iodinating serum albumin, is commercially available (Worthington Biochemical Corp.), and may be useful where avoidance of contamination of the product by the enzyme is important. The quantitative determination of iodotyrosine and iodohistidine content of proteins can be performed as described in 9; 2.2.4 and 9 2.8.2 after complete enzymic digestion. These derivatives are not stable under the usual conditions of acid hydrolysis. An alternative procedure for labelling proteins to high specific Lictivities by acylation with a 1251-containingreagent has been described by Bolton and Hunter (1973).

-

3.8. Modifications of sulfhydryl and disulfide groups The half-cystine residues of many proteins are present in the form of disulfide bridges. Some proteins contain both sulfhydryl and disulfide groups. Since sulfhydryl groups are readily oxidized, and since disulfide interchange complicates the determination of half-cystine pairing in proteins containing several disulfide bonds, it is customary Subject index p. 201

102


to convert the half-cystine residues to stable derivatives prior to undertaking chemical studies. The most generally accepted approach is the reduction of the disulfide bonds and subsequent modification of the resulting SH groups. Complete reduction of disulfide bonds in proteins, in general requires unfolding of the molecule, and, hence is performed in presence of denaturing agents such as urea, guanidine, or sodium dodecyl sulfate. The remainder of this chapter deals, in the main, with the preparation of a number of derivatives of the cysteinyl residue; such as the acidic compounds, cysteic acid, S-carboxymethylcysteine, S-sulfocysteine, and the S-sulfenylthiosulfate of cysteine, the basic derivative S-aminoethylcysteine and S-carboxamidomethylcysteine. Procedures are also presented for the application of N-ethylmaleimide, p-mercuribenzoate, 5,5’-dithiobis(2-nitrobenzoate), and azobenzene-2sulfenyl bromide to the spectrophotometric determination of sulfhydryl groups. 3.8.1. Modification by perjormic acid oxidation Treatment of proteins with performic acid leads to the oxidation of cysteine and cystine residues to cysteic acid residues (Sanger 1949). Methionine residues are quantitatively converted to the sulfone (Hirs 1956), and tryptophan undergoes oxidative destruction (Toennies and Homiller 1942; Benassi et al. 1965). Other amino acids are not modified, provided that precautions are taken to avoid chlorination (Thompson 1954; Hirs 1956), or bromination (Sanger and Thompson 1963) of tyrosine residues. Cleavage of peptide bonds does not occur on performic acid oxidation at low temperature. The procedure described by Moore (1963) is the method of choice when performic acid oxidation is employed for the quantitative determination of the half-cystine plus cysteine content of proteins. This procedure is described in detail in $2.5.1. The yield of cysteic acid is 94+ 2 % and that of methionine sulfone 100 & 2 %. A widely used procedure for the preparation of oxidized protein derivatives for amino acid sequence analysis has been described by Hirs (1967).A somewhat different procedure, used successfully in our

Ch. 3


103

laboratory for a number of years, is as follows: Performic acid is prepared by the addition of 1 ml 30% H,O, to 9 ml of formic acid (98 to 100%)at room temperature. The mixture is allowed to stand for 2 hr and is then cooled to 0°C. The reagent is then added to lyophilized protein in a precooled container, to give a final protein concentration of 17;. After 18-24 hr at O'C, the reaction mixture is diluted with an equal volume of ice-cold water, and transferred to dialysis tubing. Dialysis is performed against two changes of 100 volumes of water at O"C, and finally against 100 volumes of M mercaptoethanol, at 0°C. The protein derivative is then lyophilized. 3.8.2. Reduction and carboxymethylation (or carboxamidomethylation) The following generally applicable procedure is based on the method of Crestfield et al. (1963), and has been widely used with minor variations for the past ten years. To 5 to 100 mg of protein in a 10 ml glass-stoppered centrifuge tube, maintained under a nitrogen barrier, is added 3.61 g of deionized crystalline urea, 0.30 ml of EDTA solution (50 mg of disodium EDTA per ml), 3.0 ml of Tris buffer, pH 8.6 (5.23g of Tris and 9 ml of 1.0 N HC1 diluted to 30 ml with water), and finally 0.1 ml of mercaptoethanol. The solution is made up to a 7.5 ml mark with water, and deaerated by flushing with nitrogen for several minutes. The tube is then stoppered and allowed to stand at room temperature (22-25°C) for 4 hr, after which a freshly prepared solution of 0.268 g of recrystallized iodoacetic acid in 1.O ml of 1.O N NaOH is added. The iodoacetate used is slightly less, on a molar basis, than the amount of mercaptoethanol. The sulfhydryl groups of cysteinyl residues react most rapidly, and since the excess iodoacetate reacts faster with mercaptoethanol above pH 8 than with thioether sulfur, alkylation of methionine is kept to a minimum. The three equivalents of Tris per equivalent of iodoacetate keep the solution more alkaline than pH 8.3. Both the alkylation reaction, and the subsequent isolation of the protein derivative should be performed in the dark to prevent the formation of iodine. The protein derivative may be recovered by lyophilization, following gel filtration or exhaustive dialysis. S u h p I index p 201

104


Reduction with dithiothreitol (DTT). The use of DTT, in place of mercaptoethanol, has gained wide acceptance. The reduction proceeds essentially to completion at low levels of DTT, because of the thermodynamically favored formation of a 6-membered ring. The equilibrium constant at pH 7 and 25°C for the reduction of cystine by DTT is lo4, as compared to approximately 1 for mercaptoethanol (Konigsberg 1972). A representative procedure, patterned upon that employed for the complete reduction and alkylation of a yG immunoglobulin (Waxdal et al. 1968), is given below. A 1&20 mg/ml solution of protein in 6 M guanidine hydrochloride, 0.5 M in Tris chloride, and 0.002 M in EDTA, at pH 8.1, is placed in a glass-stoppered centrifuge tube. The solution is flushed with nitrogen, the tube closed tightly, and placed in a 50°C water-bath for 30 min to denature the protein. Dithiothreitol(50 moles/mole of disulfide in the protein, or if this is not known, 300 moles/mole of protein) is added, the tube flushed briefly with nitrogen and maintained at 50°C for 4 hr. The solution is then cooled to room temperature, and an aqueous iodoacetic acid solution (2-fold molar ratio to the DTT added previously) is added. The pH is monitored, and, if necessary maintained at 8.1 by the addition of 1 M NH,OH. After alkylation for 20 min, in the dark, the reagents are removed (also in the dark) by rapid dialysis against ice-cold water, or by passage through a column of Sephadex G-10, equilibrated with 0.01 M NH,HCO,, at pH 8.1. In both of the above procedures, iodoacetamide may be substituted on an equimolar basis for iodoacetic acid with no alteration in reaction conditions. S-Carboxymethylcysteine is readily determined by amino acid analysis following acid hydrolysis under the usual conditions (8 2.5.3). The occurrence and extent of any alkylation of lysyl, histidyl, or methionyl residues would be revealed at the same time (82.510, 2.7.1, 2.8.1). 3.8.3. Carboxyethylation of SH Free S-carboxymethylcysteine, or S-carboxymethylcysteine at the

Ch. 3

105


N-terminus of a peptide, undergoes intramolecular cyclization in acid solution to yield 3-oxo-(2H, 3H, 5H, 6H-1,4-thiazine)-5-carboxylic acid (Bradbury and Smyth 1973), which is ninhydrin-negative. CONHR

CONHR

I

/CH\ H2N HOOC

I

H\

CHZ

I

+

\

HN OC\

CHZ

CH2

I

I

+

H2O

AS CH2

Such cyclization reactions have also been reported for the Scarboxamidomethyl derivative, and for the N-ethylmaleimide adduct of cysteine. The rate and extent of this reaction depend upon the temperature and pH. S-carboxymethylcysteine cyclized most rapidly at pH 3.0 only 34 % remaining after 16 hr at 110°C in 0.2 M sodium citrate buffer. Under the usual conditions of acid hydrolysis for amino acid analysis, e.g. 6 N HCl, lIO"C, 18 hr, S-carboxymethylcysteine was stable (9 2.5.3). Bradbury and Smyth (1973) advocate the use of 3-bromopropionic acid in the place of iodoacetic acid, or its amide, in the alkylation reaction. S-carboxyethylcysteine was found to be completely stable at acid pH, and could be readily quantitated by amino acid analysis (9 2.5.3). The half-life of the reaction of the SH-group of cysteine (2 pmoles/ml) with 3-bromopropionate (200 pmoles/ml) at pH 8 and 30°C was approximately 10 min. Under the same conditions, the reaction of cysteine and iodoacetate was found to be virtually instantaneous (Bradbury and Smyth 1973). 3.8.4. Aminoetlzylation of SH CH2 R-SH

+

I

\

pH 8.9

* R-S-CH2CH2NH2

/NH

CH;

This modification is generally performed to introduce new points for tryptic cleavage within the polypeptide chain (Lindley 1956; Raftery Subtecr indexp 201

I06


and Cole 1963, 1966). Trypsin cleaves peptide bonds involving residues of S-(P-aminoethy1)cysteine which is an analogue of lysine ; cleavage is much slower than for bonds involving lysyl or arginyl residues. In principle, reduction and aminoethylation of disulfide-containing proteins. after modification of lysyl and arginyl side chains, should lead to a derivative which can be selectively cleaved by trypsin at S-aminoethylcysteinyl residues. Such a sequence of reactions has been applied to a y-globulin (Slobin and Singer 1968). Protein (25 mg) is dissolved in 5 ml of fresh 8 M urea (or 6 M guanidine HCI) in M Tris-HC1 buffer, pH 8.9, containing 75 p1 of 2-mercaptoethanol. After 2 hr at room temperature under a N, barrier, 150 pl of ethylenimine are added, and the alkylation allowed to proceed for 2 hr under N,. (The completeness of the alkylation reaction may be assessed by assaying for the absence of free thiol. One drop of the reaction mixture is mixed with a drop each of 10% w/v sodium nitroprusside solution and 1 M NH, solution. A negative reaction is indicated by absence of significant purple color.) The completely alkylated reaction mixture is dialyzed, first against 100 volumes of 1 % NH,HCO,, and then against distilled water. The protein derivative is recovered by lyophilization. Prolonged exposure of the reaction mixture to acidic conditions, prior to the removal of the excess ethylenimine, should be avoided. Significant alkylation of methionine can occur under such conditions (Schroeder et al. 1967). Some alkylation of the a-amino group may be observed (Shotton and Hartley 1973), particularly if the alkylation reaction is allowed to proceed for a prolonged time in the presence of a large excess of ethylenimine. The S-aminoethylcysteine content of the protein derivative can readily be determined following acid hydrolysis of the protein under the usual conditions, and amino acid analysis (5 2.5.4).

3.8.5.Trimethylaminoethylationof cysteine The cysteinyl residue may be converted to a strongly basic derivative, 4-thialaminine, by alkylation with (2-bromoethyl)trimethylammonium

Ch. 3

107

MODIFICATION OF PROTEIN SIDE-CHAINS n

CHZSH

I

HOOC-CH-NH,

CH3Br 119

+ CH,-N-CH,-CH,-Br I

CH3

-

n

B r CH,

CH,-S-CH,-CH I

HOOC-CH-NH,

@I

N-CH, 2- I CH,

t HBr

4-thialaminine

bromide (Itano and Robinson 1972). The virtues of this modification reaction lie in the very high stability of 4-thialaminine under conditions of acid hydrolysis, and the ready solubility of a number of trimethylaminoethylated proteins in aqueous solution. For complete alkylation of half-cystine residues, after the DTT reduction procedure described in 9 3.8.2.1, (2-bromoethyl) trimethylammonium bromide is added in a 3-fold molar excess over the DTT concentration. After reaction for 24 to 48 hr under nitrogen, at room temperature, the protein derivative is desalted either by gel filtration, or exhaustive dialysis, and lyophilized. The 4-thialaminine content of the protein derivative is determined by amino acid analysis after acid hydrolysis (8 2.5.6). 3.8.6. Reduction und selective S-methylation Conversion of half-cystine residues in proteins and peptides to the S-methyl derivatives is advantageous in subsequent studies of amino acid sequence. Under the usual conditions of acid hydrolysis (9 2.1), S-methylcysteine is recovered in a 90% yield (Heinrikson 1971). The phenylthiohydantoin of S-methylcysteine is readily identified by routine thin layer chromatography procedures (Rochat et al. 1970). With the increasing use of the sequenator, PTH-S-methylcysteine offers a marked advantage over derivatives such as PTH-cysteic acid, or PTH-carboxymethylcysteine, which have to be identified by special techniques (Edman 1960, 1970). S-methylcysteinyl residues provide a new point of cleavage for cyanogen bromide (5). The procedure described below is taken from a detailed publication o f Heinrikson (1971). Lyophilized or crystalline protein (10 to 50 mg) is dissolved in 5 ml ofa solution at pH 8.6 containing 6 M guanidine hydrochloride, 0.25 M Suhieci mder p 201

I08


Tris, 3.3 mM disodium EDTA, and 25% (v/v) of acetonitrile. (The organic solvent must be included to ensure solubility of the reagent, methyl-p-nitrobenzenesulfonate.) The solution is flushed with nitrogen, and a 50-fold molar excess of B-mercaptoethanol over the half-cystine content of the protein is added under a nitrogen barrier. The tightly closed vessel is then incubated at 50"C.After 1 hr, the reduction mixture is cooled to 37OC, and 0.5 ml of a solution of methyl-p-nitrobenzenesulfonate in acetonitrile is added under a nitrogen barrier. The amount of methylating agent added is in a 2-fold molar excess with respect to the mercaptoethanol. After incubation for 2 hr at 37"C, the clear yellow solution is treated with 5 p1 of mercaptoethanol and dialyzed exhaustively against distilled water. The protein derivative tends to precipitate during dialysis and is recovered by lyophilization. Under the above conditions, the modification is totally restricted to sulfhydryl groups (Heinrikson 1971).S-methylcysteine is quantitated by amino acid analysis after acid hydrolysis ($ 2.5.5). 3.8.7. S-Sulfonatioii In the presence of catalytic amounts of cysteine, or p-mercaptoethylamine, and oxygen, sodium sulfite converts the half-cystine and cysteine residues of proteins to the S-sulfonate derivatives. R-S-S-

t

R

so;

-

[Ol

*R-S-SOT

+

RSS

I

The S-sulfo derivative is stable at neutral and acidic pH. Sulfitolysis has been employed as a mild procedure for the cleavage of disulfide bonds in multi-chain proteins, as a preliminary to polypeptide fractionation, and also to provide a protective blocking group for the cysteinyl residue. The S-sulfo derivative is readily reduced back to the thiol by the addition of sulfhydryl compounds, e.g. /I-mercaptoethanol. The following sulfitolysis procedure is based upon the work of Chan (1968). Protein (2-5 mg/ml) is dissolved in 0.1 M Tris chloride buffer, pH 8.0, containing 8 M urea (or 6 M guanidinium chloride), 2 x M

Ch. 3


109

,&mercaptoethylamine, and 0.05 M Na,SO,. The mixture is stirred gently at room temperature for 1 hr, and then dialyzed exhaustively against distilled water. The derivative is then lyophilized. The completeness of the reaction is best assessed by employing Na2 3 5 S 0 , ,of known specific activity, and determining the incorporation of radioactivity into the protein (Chan 1968). Chan (1968) describes the following procedure for the regeneration of free thiol groups. The S-sulfonated protein (1 mg/ml) is dissolved in 1 M Tris chloride buffer (pH 7.5) containing 4 M urea, and 8-mercaptoethanol is added to a final concentration of 0.7 M. The mixture is incubated for 2 hr at 25"C, and then dialyzed against a buffer solution, appropriate to those further manipulations to which the protein is to be subjected. 3.8.8. Conversion to the S-su~enylsulfonatederivative The high reactivity of sulfhydryl groups imposes serious limitations on attempts to modify selectively other amino acid residues in proteins containing cysteinyl residues. The reversible blocking of free sulfhydryl groups in native proteins with tetrathionate offers one solution to this problem : protein-SH + S,Og

-G==

R-S-S-SO,

+ S,O5 + H i

An illustration of this approach may be seen in the studies on streptococcal proteinase (Liu 1967). The activity of this enzyme is dependent upon the presence of a free sulfhydryl group. The active form of the enzyme was first converted to the inactive S-sulfenylsulfonate derivative. Treatment of this derivative with a chemicallyreactive substrate 'analogue, a-N-bromoacetylarginine methyl ester, resulted in the alkylation of a single histidine residue. The sulfhydryl group of the modified enzyme was regenerated by reduction, however, this did not restore enzymatic activity, thus providing presumptive evidence for the involvement of both a cysteinyl and a histidyl residue in the active site of this enzyme. This method of reversible modification of sulfhydryl groups has also Suhlecr mdr.2 p Z l l i

I10


proved very useful in minimizing autolysis and proteolysis during purification of sulfhydryl proteases and, also, of other proteins from extracts containing high concentrations of sulfhydryl dependent proteolytic enzymes. Protein (1 pmole) is dissolved in 5 ml of 0.02 M sodium phosphate buffer at pH 7.0,O.OOl M in EDTA, and 35 mg of sodium tetrathionate (NaZS,0,.2H,0). After 60 min at room temperature, the reaction mixture is passed through a Sephadex G-25 column (approx. 1.5 x 40 cm) equilibrated with 0.02 M phosphate, 0.001 M in EDTA at pH 7, to remove excess reagent. The thiol groups may be regenerated by adjusting the pH to 7.6, adding a-mercaptoethanol to a concentration of 0.1 M and incubating at 37°C for 15 min. If prolonged storage of the S-sulfenylthiosulfate protein derivative is desired, this should be done in the presence of 0.001 M sodium tetrathionate. It should be noted that the formation of a stable sulfenylthiosulfate is a phenomenon observed with native proteins (Pihl and Lange 1962). When simple thiols such as cysteine are reacted with an excess of tetrathionate, S-sulfocysteine is formed (Inglis and Liu 1970). If thiol is in excess, the sulfenylsulfonate can react with a second molecule of thiol to form a disulfide. Inglis and Liu (1970)have developed a procedure for the quantitative determination of cystine and cysteine in acid hydrolyzates of proteins, dependent upon the quantitative conversion of cysteinyl residues to S-sulfocysteine, and subsequent quantitation of this derivative by amino acid analysis (9; 2.5). 3.8.9. Addition of thiols to activated double bonds 3.8.9.1. Spectrophotornetric titration with N-ethylmaleimide ( N E M ) CHZSH H~dH-COO-

CH-CO\

+

II

CH-CO’

NC2H5

-

CH2-S -CH-CO\

‘H3d

I

I

CH

CH2-CO’

N-CzHs

‘COO-

S- (ethy1succinimido)-cysteine

The spectrophotometric procedure requires freshly prepared stock solutions of NEM (1.25 x lop3 M) in either 5 M guanidine hydro-

Ch. 3


111

chloride, or in 2 % sodium dodecyl sulfate, in 0.1 M sodium phosphate buffer at pH 7.0, and a stock protein solution (approximately 5x M) in the same solvent as the NEM. The following solutions are then prepared. A : 2.0 ml NEM solution + 0.5 ml solvent B : 2.5 ml solvent C : 2.0 ml NEM solution + 0.5 ml protein solution D : 2.0 ml solvent + 0.5 ml protein solution The absorbancy (A,) of solution A (sample cell) versus solution B (referencecell),and the absorbancy (A,) of solution C uersus solution D, are then determined in 1 cm pathlength spectrophotometer cells. The reaction with denatured proteins is generally rapid and complete within 10 min under the conditions specified here. Lack of further change in absorbance over an additional period of 10 min indicates that the reaction is indeed complete. The eM at 305 nm for NEM is 620 whereas S-(ethy1succinimido)cysteine does not absorb at this wavelength. Hence, the sulfhydryl titer is given by the following relationship. sulfhydryl concentration =

AI-A~

620

rnoleshiter

3.8.9.2. Other applications of N-ethylmaleimide and related compounds Although NEM has been largely superceded as a spectrophotometric titrant for thiol groups, the reagent offers other useful features. Upon acid hydrolysis, S-(ethy1succinimido)-cysteine gives rise to S-succinylcysteine (more precisely, 2-amino-2-carboxyethyl-mercaptosuccinic acid), which is stable to acid hydrolysis and can be quantitated by amino acid analysis ($2.57).Ethylamine, the other product of the hydrolytic cleavage of the derivative,can likewise be quantitated by amino acid analysis (52.5.7). Since this product arises also from the hydrolytic cleavage of NEM itself, such quantitation is useful only after complete removal of excess reagent by exhaustive dialysis, or gel filtration, before subjecting the derivatized protein to acid hydrolysis. It should be noted that complete conversion of S-(ethy1succinimido)-cysteine to SSubject index p . 201

112


succinylcysteine requires 72 hr of hydrolysis in 6 N HCl at 110°C in uacuo (Smyth et al. 1971; Guidotti and Koningsberg 1964). Parenthetically, it should be noted that the reaction of maleic anhydride with sulfhydryl-containing proteins also leads to the formation of S-succinylcysteinederivatives, by the reaction sequence R-SH + CH-CO,

II

CH-CO

,o-

R-S-CH-CO

I

H20

'0-

CH2-COf

+

R-S-CH-COOH

I

CHzCOOH

NEM is not completely specific for sulfhydryl groups. Reaction with amino groups occurs readily in native proteins. Formation of such derivatives would be detected by a comparison of the yield of Ssuccinylcysteineand ethylamine,or by the appearance of N-~-succin-2yl-lysine, or N-succin-2-yl-histidine(Holbrook and Jeckel 1969)upon acid hydrolysis of the modified protein (cf. ch. 2). Radioactive reagents, such as 14C-ring labeled NEM (see, for example, Klee and Gladner 1972), or N(N'-a~etyl-4[~~S]-sulfamoylpheny1)maleimide (Holbrook and Pfleiderer 1965; Holbrook and Jeckel 1969) have been employed to tag specific residues in proteins and facilitate purification of peptides containing these residues. The colored derivative, N-(dimethylamino-3,5-dinitrophenyl)maleimide (DDPM) (Witter and Tuppy 1960) has been used to label cysteinyl residues in several proteins. Purification of peptides incorporating this derivative has been facilitated in several instances by their strong adsorption to talc (Witter and Tuppy 1960; Gold and Segal 1964, 1965). It should be noted that the DDPM derivative of N-acetylcysteineexhibits an of only 3000 (Gold and Segal1964), i.e. it is a relatively weak chromophore in the visible region of the spectrum. Derivatives formed by the nucleophilic addition of sulfhydrylgroups to the activated double bonds of acrylonitrile (Plummer and Hirs 1964), and 4-vinylpyridine (Cavins and Friedman 1970) have also been exploited in sequence and analytical studies.

Ch. 3


113

3.8.10. Titration of thiols with 5,5'-dithiobis( 2-nitrobenzoate) ( D T N B ; Ellman's reagent)

Following its introduction in 1959 (Ellman), this sensitive spectrophotometric procedure for the determination of thiols has gained wide acceptance. The mixed disulfide product can be exploited as an intermediate in the preparation of other useful derivatives of the cysteinyl residue. For the determination of total sulfhydryl group content, protein (0.02 to 0.05 pmole) is dissolved in 2.5 ml of 0.1 M Tris chloride buffer at pH 8.0, containing 0.01 M EDTA, and either 6 M guanidinium chloride, 8 M urea, or 1 % sodium dodecyl sulfate. A fresh solution of DTNB (0.01M in 0.05 M sodium phosphate buffer, pH 7.0) is prepared. The solvent (2.50 ml), in which the protein is to be dissolved, is placed in the reference cell of a spectrophotometer, and 2.5 ml of the protein solution in the sample cell. DTNB solution (100 pl) is added to each cell, and the absorbance at 412 nm is recorded with time. After the development of the maximum color the downward sloping line (a consequence of the spontaneous hydrolysis of DTNB) is extrapolated to zero time to determine the concentration of thiol. Where conjugated proteins, absorbing in the visible region, are under study, or where the protein contributes significant light scattering, the absorbance of the protein solution at 412 nm should be determined against a water blank, and the appropriate correction made. The values quoted for e4, of 3-carboxylato-4-nitrothiophenolate in aqueous solution have ranged from 13 600 (Ellman 1959; Janatova et al. 1968) to 14 140 M-' cm-' (Gething and Davidson 1972). The value of 13600 is that generally used. In 6 M guanidinium chloride, ~ ~ ~ ~ = 1 3M-' 8 8 cm-', 0 and in 8 M urea, ~ ~ ~ ~ = 1 M-' 4 2 9cm-' 0 (Gething and Davidson 1972). Where a determination of total half-cystine plus cysteine content of Subject index p. 201

114


proteins is desired, the protein may be reduced by any of the procedures described in Q 3.8.2, and freed of reducing agent by gel filtration on a column of Sephadex G-25, equilibrated and eluted with the solvent used for the titration. The paper by Janatova et al. (1968) should be consulted for a careful study of the optimum conditions for the determination of the sulfhydryl content of native proteins by titration with DTNB. It should be noted that incomplete conversion of protein sulfhydryl groups to the mixed disulfide leaves open the possibility of intramolecular reactions leading to the formation of disulfide bonds. protein-SH

I

(SH),

+

Ar-S-S-Ar

i protein-S-S-Ar

I

+ ArSH

W),, protein-S-S-Ar

I

===

(SH),

->s

protein - S dH,

+ ArSH

,,

where Ar =

Qcoo-

Such interchange reactions have been observed with a number of proteins, and are particularly likely to occur where initial reaction of the most rapidly reacting sulfhydryl groups on a protein molecule is followed by removal of excess DTNB, and a denaturing treatment which leads to the exposure of previously unavailable thiol groups. 3.8.1 1 Cyanylation of S H

Degani et al. (1970) and Degani and Patchornik (1974) demonstrated that quantitative cyanylation of sulfhydryl groups in proteins may be achieved by reaction with 2-nitro-5-thiocyanobenzoate (NTCB).

FI-SH+

6 - bco0: coo-

NO2

p~ 7-a

RSCN +

1

NO2

Ch. 3

115


The preparation of NTCB has been described by Degani and Patchornik (1971). To a 150 ml aqueous solution containing 7.5 g of K H C 0 3 and 2.0 g (31 mmol) of KCN, 3.0g (7.5mmol) of 5,S-dithiobis(2-nitrobenzoic acid) were added with stirring. After 30 min, a freshly prepared 3 solution of CNBr in water was added slowly (10 min) to the stirred deep orange solution until the color was completely discharged. This required 27 ml. The pH was then decreased to 5 by the dropwise addition of glacial acetic acid, excess cyanide was removed by bubbling a stream of N, through the solution for 12 hr. Upon acidification to pH 2.3 with 6 N HC1 a white solid crystallized out. After cooling the mixture with ice, the solid was filtered, washed with cold water and air-dried, yield 3.42 g (94 %) of chromatographically and electrophoretically pure NTCB, mp 248°C.

@O0S I S

CNPH 8

.

SCN

S-

Qco0No2

For complete cyanylation of sulfhydryl groups, protein (1 x M, or less than 10 mg per ml; concentrations higher than this should be avoided) is dissolved at room temperature in 0.1 M phosphate, pH 7.4, containing 6 M guanidine HC1 and 1 mM EDTA, and the solution freed of oxygen by flushing with N,. A5-fold molar excess of NTCB in the same buffer is then added and the rate of reaction followed by monitoring the absorbance at 412 nm, due to the 2-nitro-5-thiobenzoate anion ( E ~ =13600). When the reaction is complete, the protein is separated from the reagents by either dialysis or gel filtration at nearneutral pH. If conversion of both sulfhydryl and disulfide groups to the thioSubjecr index p. 201

116

CHEMICAL MODIFTCATION OF PROTEINS

cyanate derivative is desired, the protein should first be reduced with dithiothreitol in 6 M guanidine HCl (fi 3.8.2.1)and the reduced protein separated from the reducing agent either by gel filtration or dialysis prior to being subjected to the above cyanylation procedure. NTCB will undoubtedly prove to be a very popular reagent for the modification of sulfhydryl groups. The reagent appears to be totally specific. The -CN substituent on the -SH group is small in comparison with that introduced by reaction with 5,5’-dithiobis(2nitrobenzoate), p-mercuribenzoate, N-ethylmaleimide or indeed any of the other commonly used -SH reagents. Hence, this reagent may permit clarification of situations where loss of biological activity due to reaction with sulfhydryl groups results from steric hindrance due to the substituent rather than to the substitution of a functionally vital thiol group. Another advantage is that labeled reagent is readily synthesized by using Na14CN. Selectivecleavage of the polypeptide chain at the thiocyanatoalanine residues may be achieved by incubation at alkaline p H , e.g. p H 9.0 for 24 hr at 37°C. Under these conditions some conversion to an uncleaved dehydroalanine peptide also occurs. R-CO- NH-CH-CO-

NH R’

I CHZ I NGC-S

pH 9.0, 37OC +

R-COOH + HN-CH-CO-NH-R‘

major reaction

I

I 1

I

1

HN=C

CHZ

‘S’ pH 9.0. 37OC minor reaction

R-CO-NH-C-CO-NHR’ I1 CH2

+

SCN-

A detailed account of the procedure for the specific cleavage at the amino peptide bonds of cysteine and cystine residues of several proteins, utilizing NTCB, has been presented by Jacobson et al. (1973). 3.8.1 2. Reaction of thiols with organomercurials 3.8.1 2.1. Titration with p-mercuribenzoate ( P M B ) Reagent grade sodium p-chloromercuribenzoate (9 mg) is dissolved in

Ch. 3

MODIFICATION OF P R ~ T E I NSIDE-CHAINS

117

1 ml of 0.04 M NaOH, the solution is made up to 25 ml with water, and centrifuged to remove particulate material. This solution (approxiM PMB) should be stored at room temperature in the mately 8 x dark. Prior to use, the exact concentration of the reagent is determined by measuring the absorbance of diluted solutions obtained by adding 0.5 ml of the stock solution of PMB to 9.5 ml of 0.33 M sodium acetate buffer at pH 4.6 ( E = 1.74 x lo4 M - ' cm-' at 234 nm). For an initial determination on a protein of unknown thiol content, it is appropriate to use a stock solution of protein at a concentration of 1 mg per ml in 0.33 M sodium acetate at pH 4.6, containing either 8 M deionized urea, or 2 % sodium dodecyl sulfate. This solution should be centrifuged prior to use. (At this protein concentration, a AA at 255 nm of approximately 0.2 would be observed upon titration of a macromolecule of molecular weight of 30 000 containing a single cysteinyl residue). Precisely measured aliquots (10 pl) of stock PMB solution are then delivered to a reference 1 cm light path silica cell containing 3.0 ml of solvent and to a matching cell containing 3.0 ml of protein solution. The contents of the cells are mixed after each addition (very small Teflon coated magnetic stirring bars are very useful for spectrophotometric titrations), and the change in absorbance at 255 nm is recorded. The titration is continued until no further change in absorbance is observed. The equivalence point is determined graphically (see Fig. 3.2), and the thiol content computed from the known concentrations of the reagent and the protein. The AcM at 255 nm for PMB-cysteinyl residues is in the range of 6000-8000 M - ' cm-' at pH 4.6, depending on the solvent. An excellent detailed discussion of the application of PMB to the determination of sulfhydryl groups in native proteins has been presented by Benesch and Benesch (1962). 3.8.12.2. Reaction with mercurated nitrophenols McMurray and Trentham (1969) and Stefanini et al. (1972) have described the synthesis of several mercurated nitrophenols (Table 3.2). The spectra of these organomercurials show large changes in the visible region when thiols displace more weakly bound ligands from Subject index p. 201

118


TABLE3.2 Spectral properties of mercurated nitrophenols* Compound

Structure

LmBx

mfl

&(in0.1 M KOH) M - cm -

2-chloromercuri4-nitrophenol

2-chloromercuri4.6-dinitrophenol

4-chloromercuri2-nitrophenol I

HgCl

2,6-dichloromercuri4-nitrophenol

2-acetoxymercuri3-nitrophenol

* Data from McMurray and Trentham (1969), and Stefanini el al. (1972). ** In 0.1 M NaOH.

119

0.8

:

Lo Lo

0.6 c 0

0 0 c

2 0.4 L

In 0

n

a 0.2 A

0

I

I

I

I

I

I

I

I

20 40 60 80 p - C h l o r o - r n e r c u r i b e n z o a t e (UP)

Fig. 3.2. Titration of a SH-group containing protein with p-chloromercuribenzoate (PMB). A, is the absorbance of the protein solution prior to the addition of PMB. The increment in absorbance, B, due to mercaptide formation, observed upon titration with PMB (as described in the text), is a measure of thiol concentration in the solution.

the mercury atom. These spectral changes are very largely a consequence of the perturbation of the pK of the nitrophenols. These compounds have been used as probes of the environment of sulfhydryl groups in native proteins, as well as in the preparation of crystalline mercurial derivatives of proteins for X-ray diffraction analysis. Aside from these applications, the mercurated nitrophenol derivatives may be used for the spectrophotometric titration of sulfhydryl groups in the same manner as PMB, with the added advantage that the spectral shifts associated with mercaptide formation may be followed in the visible region of the spectrum where interference from protein absorbance is minimal. 3.8.13. Reaction of cysteinyl residues with azobenzene-2-sulfenyl bromide Azobenzene-2-sulfenyl bromide is reported to react selectively with Subject indexp. 201

120

CHEMICAL MODlPlCATlON OF PROTEINS

cysteinyl residues in proteins. The reagent is soluble in water and reacts readily with thiols in acid solution (pH 1-5). 8 Br

The following procedure, applied to egg-white lysozyme, was described by Fontana et al. (1968). Protein (40 mg) was dissolved in 2 ml8 M urea solution, adjusted to pH 8.6 by the addition of 5 % methylamine, and P-mercaptoethanol (20 pl) added. After 1 hr at room temperature, the protein was precipitated with acetone: 1 N HCl(39 :1, v/v) at - 592, separated by centrifugation, and washed with acetone. The protein was dissolved in 2 ml of 25 o/, (v/v) aqueous acetic acid. Azobenzene-2-sulfenyl bromide (35 mg) was dissolved in 50% aqueous acetic acid (1 ml), and added to the solution of the reduced protein. After 30 min at 22-24"C, the protein derivative was freed of excess reagent by passage through a column of Sephadex G-25 (90 x 1 cm), equilibrated and developed with 0.2 M acetic acid, and lyophilized. The modified protein was dissolved in 0.1 N HCl (2 mg/ml) and an aliquot (0.3ml) added to 2.7 ml of 8 M urea adjusted with HCl to pH 1. The concentration of azobenzene-2-sulfenyl residues ( E = 16 700 at 353 nm) was determined spectrophotometrically, and extent of modification calculated by comparing this concentration with the concentration of protein in the 0.1 N HCl solution, determined by amino acid analysis on an appropriate aliquot. The mixed disulfide is readily reduced by thiols, such as P-mercaptoethanol, or DTT, or by sodium borohydride.

CHAPTER 4

Site-specific modification of native proteins with group-specific reagents

The general objectives for the site-specific modification of native proteins include: (1) the identification of amino acid residues at the catalytic site of enzymes and the ligand-binding sites of regulatory proteins, and of immunoglobulins as well as enzymes; (2) the introduction of physico-chemical reporter groups such as fluorescent, spectrophotometric or spin-label probes; and (3) the labelling of a specific protein in order to isolate it from a multicomponent system such as a ribosome or mitochondrion. The potential use of irreversible inhibitors as pharmacological agents provides another motivation for their study. As the term will be used here, site-specific modification implies either the exclusive derivatization of a single amino acid residue or the incorporation of a single mole of modification reagent per mole of protein subunit. The site-specific modification of native proteins is not one of the routine procedures in protein chemistry. It cannot be placed in the same category as end-group labelling or determination of amino acid composition and sequence. The specific chemical modification of a native protein can never be guaranteed because the reactivity of amino acids in a native protein is rarely predictable even if the threedimensional structure of the protein is known. Unusual pK,’s of side chains, steric and solvent effects and the proximity of the amino acid residue to a ligand-binding site all influence its reactivity, frequently in opposite directions. However certain well-defined avenues of in121

S u b p i mdcr p 201

122


vestigation are available which the skillful experimenter may follow to answer his specific question. In general, there are two approaches to the site-specific modification of a protein: side chain specific modification reagents such as those discussed in ch. 3 and affinity labels or active site-directed irreversible inhibitors. The latter type of compound shares important structural similarities to ligands known to have affinity for a particular binding site but is distinguished from reversible inhibitors by the possession of reactive groups capable of forming covalent bonds with amino acid side chains. Most commonly, the reactive groups are alkylating agents such as haloketones or haloacids although acylating and arylating groups have been employed. New types of reactive substituents currently receiving active attention are photoactivatable or capable of being transformed into a reactive group by the catalytic site of an enzyme. Many side-chain specific reagents have been successfully used for the stoichiometric modification of native proteins. This approach accounts for most of the papers published in this area. Although the likelihood of a successful site specific modification with these reagents is significantly less than that with a well designed affinity label, many more experiments with these compounds are probably attempted because minimal commitment of resources is necessary to initiate such studies. Most of the site-specific reagents are readily available, whereas affinity labels usually must be synthesized, often by difficult routes. The unexpected specificity which can be achieved with functional group modification reagents is an apparent consequence of the native protein’s ability to impose a unique chemical environment on a given amino acid under a given set of experimental conditions. It is important to emphasize that the site-specific modification of a protein is a kinetic phenomenon and selective modifications result from the ability of the protein to alter the reaction rate of a single residue under one clearly defined condition of pH, ionic strength and temperature. For example, it is entirely possible that if, at pH 7.0, one lysine residue is substantially more reactive than either free lysine or other lysine residues in the protein it may well be less reactive than these at pH 9.0.

Ch. 4

SITE-SPECIFIC MODIFICATION OF NATIVE PROTEINS

123

4 . 1 . Effect of complex formation Two factors are chiefly, but not exclusively,responsible for the fact that, under certain conditions, amino acids in native proteins react more rapidly than free amino acids in solution. The first and most general is the capacity of proteins to bind modification reagents at or near the functional groups of amino acid residues in orientations favorable to reaction. The reversible binary complexes formed between proteins and modification reagents prior to reaction are analogous to enzymesubstrate complexes. As a result, most site-specific modifications of native proteins probably proceed by the scheme summarized in eq. (4.1) rather than by the strictly bimolecular scheme summarized in eq. (4.2) where P' designates the derivatized protein and R is the modification reagent. ki

k2

P+RkP-R--tP' k-

1

h

P+R-+P' These two mechanistic alternatives can often be distinguished kinetically. The pseudo first-order rate constant for the modification of the protein (kobs)for the schemes summarized in eqs. (4.1), (4.2) are given in eqs. (4.3), (4.4), respectively, where the

modification is carried out under conditions in which the reagent is present in large excess relative to the protein and where in eq. (4.1)k, is assumed to be small relative to k- If the scheme involving the formation of the binary cpmplex (eq. 4.1) is valid, then the first order rate constant for modification will not always be linearly dependent on the concentration of the modification reagent. The simplest procedure to determine if intermediate complexes form during a modification Subject inder p. 201

124


reaction is to measure the pseudo first-order rate constant as a function of reagent concentration and then plot the reciprocal of this rate constant versus the reciprocal of the modification reagent concentration. Since eqs. ( 4 4 , (4.6) are the reciprocal forms of eqs. (4.3), (4.4) respectively, a finite intercept in these double reciprocal plots indicates that an intermediate complex is probably formed. The ratio of the slope to the intercept in the double reciprocal plots yields K R , the dissociation constant of the protein-modification reagent binary complex.

Failure to detect a complex by this kinetic procedure is not proof that it does not form. If the concentration of the modification reagent is substantially less than K,, then eq. (4.3) reduces to eq. (4.7) - an expression which is indistinguishable from a strictly bimolecular mechanism.

One alternative approach for demonstrating the existence of a complex during the course of a modification reaction is stereochemical in nature. For example, if enantiomers of a modification reagent give either different rates of modification or different products, the importance of multiple sites of interaction between the reagent and the protein, and hence intermediate complex formation is indicated. Examples of studies of this type include the alkylation of bovine pancreatic ribonuclease and papain by a variety of haloacids (Heinrikson et al. 1965; Eisele and Wallenfels 1968). The simple numerical analysis below indicates the significant rate acceleration which can be achieved if a modification proceeds through the formation of a reversible complex (eq. 4.1) rather than through a

Ch. 4


125

strictly bimolecular reaction (eq. 4.2) as would be the case for a free amino acid in solution. Rates of intramolecular reactions are usually compared to those of intermolecular reactions by determining the ratio of the first-order rate constants to the second order rate constants. The units of this ratio are concentration and the ratio is at least 5 M for simple organic reactions if no specific orientation effects are involved. Ratios of this size are generally interpreted in terms of the increased local concentration of one reactive group relative to another in the intramolecular reaction. If one considers that the ratio of k , , the first order rate constant of eq. (4.1), to kR, the second order rate constant of eq. (4.2), is 5 M, the expression in eq. (4.8) permits the comparison of the observed pseudo first-order rate constant for modification according to eq. (4.1) (k:bY) and eq. (4.2) (k2bs). The reaction which proceeds according to eq. (4.2) can be considered the rate of modification of the free amino acid. k2R 5 kR(R) 5 k,(10-2 M) _ _ _ _ _ _ kAb5 R + K R R + K R 1 ---=--- 0 - 2 M + 0 . S M = 10 (4.8) kts k,(R) k,(R) k , W 2 M) Assuming that the afinity of the modification reagent for the protein is slight so that K , is 0.5 M and that the modification reagent concentration is lo-* M, it is readily apparent that the rate of modification of an amino acid in a native protein is ten times greater than that for a free amino acid in solution. This conclusion is remarkable for it shows that, without assuming highly specific geometric constraints or a marked affinity of the reagent for the protein, it is still possible to achieve a significant rate enhancement if a complex is formed. Since site-specific modification with a side-chain specific reagent is exclusively a kinetic phenomenon this analysis shows one reason why amino acids of the same type on a given protein can have widely disparate reactivities. Related to the ability of the protein to interact with modification reagents and thus form binary complexes, is its probable role of providing a reaction matrix which can preferentially stabilize the Suhjecr m C r p 201

126


transition state of a particular modification reaction. The effect is roughly equivalent to the capacity of an enzyme to coordinate or chelate the transition state of the particular reaction that it catalyzes. Clearly,a number of specific interactions between protein, modification reagent and reactive amino acids are.responsible for the stabilization of the transition state. They would include electrostatic effects, hydrogen-bonding interactions and unique microenvironments of varying dielectric constants. If the transition state of the modification reaction is stabilized by additional interactions with the protein, then the selectivity of the modification reaction will be considerably enhanced. The unusual rate enhancements observed for the reaction of some functional group reagents with specific amino acids undoubtedly are explicable in these terms. Perhaps the most outstanding example is the reaction of serine esterases with diisopropyl fluorophosphate (Sanger, 1963).

4.2. Effects of p K , perturbations The perturbation of the acid dissociation constant of an amino acid residue as a consequence of its environment within a protein represents another mechanism for enhancing its reactivity relative to a free amino acid in solution. Most amino acid residues react with their respective modification reagents in their unprotonated form instead of in their conjugate acid form. Eq. (4.10) describes the pH dependence of the simple bimolecular reaction (eq. 4.9) of the free base form of nucleophilic amino acid side chain with a non-ionizable modification reagent where K , is the acid dissociation constant and A , is the total. concentration of amino acid. kR

A+R+A-R

(4.9) (4.10)

If a pH-dependent rate constant (kR')is defined by eq. (4.1l), (4.1 1)

Ch. 4


127

it is apparent that the observed rate constant at any pH will be a function of the reactivity of the nucleophile (the magnitude of k R ) and the fraction of the nucleophile in its deprotonated form [aA= K,/(H+ Ka)]. At any given pH, kk will increase with increase in K , . In effect, by increasing the acid dissociation constant (or decreasing the pK,) the protein generates a larger concentration of the nucleophilic form of the amino acid residue. However, the bimolecular rate constant changes as a function of the pK, of the nucleophilic group according to the Bronsted relation (eq. 4.12) where y is a reaction constant which depends on the nature of the modification reaction and p describes the sensitivity of a series of nucleophiles on the pK,.

+

log kR = PpK,

+y

(4.12)

This standard free energy relation indicates that the rate acceleration gained as consequence of increasing the reactive form of the nucleophile is partially offset by the decrease of k R . But since the values of are often close to 0.5, a net rate acceleration usually results especially if the pH of the reaction mixture is near the pK, of the ionizing nucleophile. The exceptional reactivities of lysine residues in glutamate dehydrogenase and acetoacetate decarboxylase with pyridoxal and 2,4-dinitrophenyl acetate respectively are almost certainly due to the greatly perturbed pK,’s of the modified residues (Piszkiewicz and Smith 1971; O’Leary and Westheimer 1968; Kokesh and Westheimer 1971). The pK, of the e-NH, group of reactive lysine in glutamate dehydrogenase is 7.9 while that in acetoacetate decarboxylase is 6. The normal pK, of lysine is 10.2 Recently, Kaplan et al. (1971) devised a new procedure for determining the ionization constants and reactivities of individual amino group of proteins with a competitive labelling technique. This method represents the first comprehensive attempt to relate the reactivity of amino acid side chains to their pKa’s and nucleophilicities. The method consists of reacting the protein and a free amino acid standard with a limiting amount of a radioactive group-specific reagent such as acetic anhydride. The protein, which is initially heterogeneously Subjecr index p. 201

128


labelled with the radioactive modification reagent, is then exhaustively modified with non-radioactive reagent to yield a chemically homogeneous but radiochemically heterogeneous final product. The internal standard (phenylalanine was used in the published studies with elastase) is present when the acetylation is performed with the non-radioactive as well as with the radioactive acetic anhydride. After digestion of the modified protein and isolation of the radioactive peptides, the specific activities of these peptides and of the acetylated amino acid standard are determined. This procedure is repeated at several pH’s

‘*‘

= cc,

specific radioactivity of acetylated peptide (4.13) specific radioactivity ofacetylated internal standard

Since the pK, of the internal standard is known, the product aAr in eq. (4.13) can be determined at any pH where and cl, are fractions of a particular lysine residue and internal standard present in the deprotonated form, respectively, (i.e. aAis defined as in eq. 4.1 1); and Y is the quotient kXy’-’/kS, where k;’-‘ is the second order rate constant for the reaction of acetic anhydride with the unprotonated form of lysine residue X on the protein and k i is the comparable constant for the reaction ofthe internal standard with acetic anhydride. Experiments at varying pH values permit the determination of the acid dissociation constant of the particular lysine residue under study. The results obtained for porcine elastase have revealed that the &-aminogroups of both lysines 224 and 87 have normal pK,’s (about 10.3) and nucleophilicities anticipated for primary amines of this pK,. That is, the second order rate constants for the reaction of the unprotonated &-amino groups of these lysine residues with acetic anhydride fit approximately on the Bronsted plot for the reaction of acetic anhydride with a variety of amines. The results for valine-16, which is known to be in a salt linkage in the interior of the enzyme, indicate that this a-amino group has an apparent pK, of 9.7 but a nucleophilicity of only 4% of that anticipated from the Bronsted plot for an amine ofthisp& Theunusuallylow reactivityof thisvalineisconsistent with the significant steric hindrance that would be expected from a resi-

Ch. 4


120

due involved in a salt linkage and shielded from the solvent.Although no highly reactive lysine residues were present in porcine elastase, the competitive labelling techniques of Kaplan et al. (1971), in principle, provide a method to determine if a site-specific modification reaction at a given pH is unusually rapid due to the unusual nucleophilicity of this residue or to its unusually low pK,. Exceptional nucleophilicity or reactivity of the residue is likely to be the consequence of the formation of preliminary binary complex of reagent and protein or of the stabilization by the native protein of the transition state for the modification reaction. In summary, the ability of the protein to bind modification reagents both prior to reaction and/or in the transition state, coupled with its capacity to increase the reaction rate by perturbing the pK,’s of the nucleophiles can account for the selectivity of a modification reagent with no apparent affinity for the protein modified. Clearly, in certain reactions, only one of these sources of rate enhancement is operative. The ability to rationalize the selectivity which can be achieved with modification reagents specific for amino acid side-chains should not be confused with the capacity to predict the results of the modification of any given protein with a particular modification reagent. For this reason, this experimental approach remains highly empirical and the ease of attempting experiments of this type should be weighed against the low probability of their success and their low information content. Clearly in the absence of sequence information and a precise understanding of the protein’s functional role and properties, an isolated specilic modification of a protein has very little intrinsic importance. Since the experimenter possesses very few means of controlling the course of a modification reaction with functional group-specific reagents, the use of affinity labels constitutes the most rational approach for the site-specific modification of a protein. It is the only method which can be used with any hope of success in a complex protein mixture.

Subjecr index p 2001

130


4.3. Widely employed site-specific reagents Several reagents which do not bear a strong structural homology to natural ligands have proved generally useful in either classifying enzymes as to basic mechanism or modifying specific types of binding sites. For example, diisopropyl fluorophosphate has played an historically important role in defining the class of serine esterases and proteases (Sanger 1963).Pyridoxal phosphate has been extremely useful in the modification of lysine residues at or near the binding site for inorganic phosphate or phosphate esters in various enzymes. Procedures involving these two generally useful site-specific reagents are now presented. 4.3.1. Diisopropvl fluorophosphate Clearly all experiments performed with the nerve gas, diisopropyl fluorophosphate (DFP), must be done in a fume hood with the utmost care. DFP rapidly hydrolyzes in 1 N NaOH, so that all glassware which comes in contact with the nerve gas should be soaked in this alkaline solution for at least 1 hr as a precaution prior to its reuse. (Warning: D F P is a potent poison. Atropine is an effective antagonist. For details of treatment, see Goodman, L. S. and Gilman, A., 1965, ‘The Pharmacological Basis of Therapeutics,’ The Macmillan Co., New York, 3rd Ed., p. 454). Stock solutions of DFP can be conveniently prepared in isopropanol in concentrations from 0.1 M to 0.001 M. These solutions are stable for a month in the refrigerator (Jansen et al. 1949). Moon et al. (1965), in their studies of the reaction of chymotrypsin with DFP, have determined the normality of stock solution of DFP in the following manner. Roughly 0.03 moles of DFP were added to an aqueous solution of 0.17 moles of KOH in a volumetric flask. The solution was allowed to stand for 12 hr at 25°C to permit complete hydrolysis, and was titrated to pH 7.0 with 0.1 N HCI. Identical titrimetric results were obtained after 43 hr standing in alkali indicating that complete hydrolysis had taken place after 12 hr. To determine the amount of free acid, if any, in the DFP, the same amount of D F P that was used

Ch. 4


131

in the basic hydrolysis is added to water and then titrated with KOH to pH 7.0. The amount of D F P in the stock solution can then be readily calculated from the two sets of titrimetric data. The derivatization of an enzyme by D F P is accomplished simply by incubating D F P with the protein. Aliquots of the incubation mixture can then be assayed for activity in the presence of D F P or after its removal by either gel filtration or dialysis. The second-order rate constant at neutral p H for the inactivation of chymotrypsin by D F P is roughly 3 x lo6 M - ' sec-' (Main 1964).This is the most rapid rate of inactivation by D F P so far observed. The rate of inactivation ofacetylcholinesterase is an order of magnitude slower (Froede and Wilson 1971). Although D F P has been and still is widely used to test if an enzyme possesses an unusually reactive serine, it should be noted that phenylmethane sulfonyl fluoride can also be used for the same purpose (0 5.7). This reagent is in some ways preferable as it is non-volatile and relatively non-toxic. Therefore, it can be employed with fewer precautions.

4.3.2. Pyridoxal-5-phosphate Several proteins are readily derivatized by pyridoxal-5-phosphate but react only slowly with pyridoxal. They include fructose-1,6-diphosphatase (Marcus and Herbert 1968), glutamate dehydrogenase (Piszkiewicz et al. 1970), glyceraldehyde-3-phosphate dehydrogenase (Ronchi et al. 1969), hemoglobin (Benesch et al. 1972), phosphofructokinase(Uyeda 1969),6-phosphogluconate dehydrogenase (Rippa et al. 1967),and phosphoglucose isomerase (Schnackerz and Noltmann 1971). Since most, if not all, these proteins possess specific phosphate binding sites, the best explanation for the effectiveness of pyridoxal phosphate as opposed to pyridoxal as a modification reagent is that it preferentially binds to these sites. Generally, pyridoxal-5-phosphate inhibits these enzymes by the formation of Schiff base with a reactive lysine residue. However, since this type of linkage is readily hydrolyzed, irreversible inhibition usually can be obtained only if the Schiff base is reduced by sodium borohydride. Subjecf index p . 201

132


In their studies with 6-phosphogluconate dehydrogenase, Rippa et al. (1967) incubated the enzyme with concentrations of pyridoxal phosphate ranging from lo-' to M for 15 min at pH 7.5 in 0.01 M phosphate buffer. After Wold dilution the mixture was incubated a further 5 min. Aliquots were then assayed for activity in standard substrate solutions which contained a concentration of pyridoxal phosphate equal to that in the second incubation mixture. The additional pyridoxal phosphate was added to prevent reversal of the reaction. Piszkiewicz and Smith (1971) obtained significant amounts of inhibition simply by withdrawing aliquots of enzyme from its incubation mixture with pyridoxal-5-phosphate without including any additional pyridoxal phosphate in the assay mixture. Clearly, the ease of reversibility of the pyridoxal phosphate inhibition will vary from enzyme to enzyme and both procedures should be employed in exploratory studies. The site of attack of pyridoxal-5-phosphate can be conveniently explored by determining which substrate and/or coenzyme protects the protein from derivatization by pyridoxal-5phosphate. For example, Rippa et al. (1967) have shown that 6phosphogluconate and inorganic phosphate protected 6-phosphogluconate dehydrogenase from inhibition while NADP' had no effect. The latter finding indicates that phosphate buffers may not be ideal for use in studies with pyridoxal-5-phosphate. At the very least, the rate of inactivation by pyridoxal-5-phosphate should be measured as a function of phosphate concentration to determine if phosphate inhibits the modification reaction. The borohydride reduction of Schiff base formed between a protein and pyridoxal phosphate should be carried out at a mildly acidic pH. Although sodium borohydride is more unstable in acidic solution, this disadvantage is offset by the exceptional reactivity of the Schiff base salts which are formed in mildly acidic solution (pH 4.5-6.5) (Schellenberg 1963). Reductions have been carried out after the protein and pyridoxal-5-phosphate have been incubated at a pH of 7.5 which is then changed to 4.5 or 6.5 with acetic acid (Rippa et al. 1967; Dempsey and Christensen 1962;Piszkiewicz et al. 1970)or after initial incubation at pH 6.0 (Schnackerz and Noltmann 1971).The relative merit ofeither

Ch. 4

133


procedure depends on whether or not the Schiff base formed with the given protein is freely reversible. Prior to reduction with sodium borohydride, octyl alcohol can be added to avoid foaming. Then a solution of 0.054.06 M sodium borohydride in either water or 0.001 N sodium hydroxide is added in equal aliquots until roughly a 100-fold molar excess has been added. After each addition, the pH of the reaction mixture should be readjusted to the initial acidic pH by the addition of either acetic acid or hydrochloric acid. The temperatures of the reaction mixture have ranged from 425°C. The modified protein can then be isolated either by precipitation, dialysis or gel filtration under conditions where the native protein is normally stable. Proof that a lysine residue has been modified can be readily obtained because pyridoxyl derivatives of lysine possess characteristic white-blue fluorescence (Ronchi et al. 1969). In addition, they have a distinctive absorption maximum at 325 nm with cM of 9710 M-' cm-' (Fisher et al. 1963). Finally, a radiochemical label can be introduced by reducing the pyridoxal-5-phosphate protein complex with tritiumlabelled sodium borohydride. The peptide containing the derivatized lysine can therefore be detected either by fluorimetry, spectrophotometry or radiochemical techniques following routine procedures of proteolytic digestion and fractionation. Acid hydrolysis in 6 N HCl for 24 hr of peptides containing pyridoxal-5-phosphate lysine yields pyridoxyl-lysine since phosphate esters are readily hydrolyzed under these conditions. Pyridoxyl-lysine is eluted between lysine and histidine from a 55 cm column of Beckman 50 resin with 0.15 M citrate buffer pH 5.28. Authentic pyridoxyl-lysine has been prepared by Schnackerz and Noltmann (1971) by borohydride reduction of a mixture of po1y-Llysine hydrochloride (20 mg) and 0.38 moles of pyridoxal, which had been allowed to stand for 10 min at 0°C in 50 mM sodium phosphate (pH 6.0). Low molecular weight impurities were removed by dialysis against 50 mM sodium acetate buffer (pH 6.0). After dialysis the resulting product can be hydrolyzed in 6 N HCl to yield pyridoxyllysine and some lysine. The latter contaminant can be minimized by \ub,
2OOO dalton) and this can also be used for concentrating. A quicker alternative to dialysis is gel filtration which is available for molecular weight down to 1000 dalton. This produces a very efficient removal of salts and other low molecular weight substances such as urea but dilutes the solution to some extent and is difficult to apply to very large volumes (greater than a few hundred ml) or to viscous solutions.Very small volumes of solution can be conveniently desalted by paper chromatography. Volatile salts can be removed by rotary evaporation or freeze drying but these methods are not recommended unless the solution must be concentrated as well. Solutions of oligonucleotides, particularly small ones, can be concentrated by column chromatography on DEAE-cellulose or charcoal. DEAE-cellulose is much more widely used. The concentrated solution contains 1 or 2 M ammonium bicarbonate or other 295

Subject index p 487

296

SEPARATION METHODS FOR NUCLEIC ACIDS

volatile substances which must be removed by repeated dilution and evaporation or freeze drying of the solution. This is an important disadvantage of the method and it is probably best used only for oligonucleotides up to 3 nucleotides long. Larger oligonucleotide solutions can be desalted on DEAE-cellulose and the desalted solution concentrated by rotary evaporation, freeze drying, dialysis against Ficoll or gel filtration in a basket centrifuge depending on what apparatus is available. These methods are described below.

6.1. DEAE-cellulose The classical method of desalting and concentrating oligonucleotide solutions consists of absorbing the oligonucleotides onto a small (1 x 10 cm) DEAE-cellulose column, washing off the salts and urea, and eluting the oligonucleotides with ammonium bicarbonate. The ammonium bicarbonate solution is then removed by rotary evaporation or freeze drying( Rushizky and Sober 1962a). This method must be applied with care. The resin should be well packed in a siliconised column and the flow rate should be 0.5-1.0 ml/min. The column should be equilibrated with dilute (1-10 mM) ammonium bicarbonate buffer pH 8.5 which will probably need degassing. The oligonucleotide solution will need diluting before loading and this may be done with 3 or more volumes water or 7 M urea. The column is loaded and then washed with 10 column volumes of water or 10 mM ammonium bicarbonate. The nucleotide material is eluted with 1 M ammonium bicarbonate, degassed, pH 8.5. Care should be taken to see that the nucleotide material does indeed absorb during loading and is not washed off during washing (Matthews 1968a).

6.2. Evaporation Evaporation of the concentrated ammonium bicarbonate solution is tedious as usually the residue must be redissolved and reevaporated

Ch. 6

DESALTING AND CONCENTRATING

291

several times to remove completely the ammonium bicarbonate. Freeze-drying suffers from the same disadvantage and it is also advisable to use only a partial vacuum to prevent oligonucleotide material being sucked into the pump. If small quantities are involved carrier nucleotides should be added and the evaporating or drying vessels should be siliconised (9 4.3.3). Triethylamine bicarbonate may be used instead of ammonium bicarbonate.

6.3. Activated charcoal Adsorption onto activated charcoal (Norit A) has been used for small oligonucleotides, up to tetranucleotides (Crane and Lipman 1953; Mandeles and Kammen 1966). The method is quicker and more convenient than DEAE-cellulose chromatography. Mandeles and Kammen (1966) used Norit A suspended in 1 mM phosphate 1 mM pyrophosphate buffer at pH 6.0 at a concentration of 100 mg/ml. The nucleotide solution is adjusted to pH 3 and 5 mg Norit for every optical density unit (-40 pg) of nucleotides is added. The charcoal is collected on a filter paper and washed with water. (A good quantitative measure of 32Por I4C radio-activity may be obtained by counting the dry filter paper in a Geiger-Muller thin window counter.) The nucleotide material is eluted with a small volume of a mixture of ethanol, water and ammonia (600:400:6.5; V/V/V) which is then removed by drying under reduced pressure at 40°C. Zachau et al. (1966a) used charcoal columns (Carboraffin C treated with 95% ethanol - 8% see. octanol, dried and degassed) at 4°C. The oligonucleotides were adsorbed, the column washed with water, and the oligonucleotides eluted with 8% aqueous phenol, pH 7. The phenol was removed with ether extraction. However, in later experiments they preferred dialysis at least for penta- and higher oligonucleotides.

6.4. Paper chromatography Small volumes may be conveniently desalted by paper chromatoSahlecr index p 4x7

298


+

graphy. The solution, 50 pl/cm is loaded as a small spot or line on a strip of filter paper. The salts are eluted with a solvent in which the mobility of the oligonucleotide material is zero such as 95% ethanol. The oligonucleotidesare then eluted with water. Very small volumes of eluent can be dried for concentration or storage under a partial vacuum as spots on a sheet of polythene or Parafilm.

6.5. Dialysis This is an extremely valuable method for large molecules. In an experiment with the hexanucleotide ( n = 6) peak from a pancreatic ribonuclease digest of p2 phage RNA it was found that 1% of the nucleotide radio-activity was lost through the dialysis bag per hour at 4°C with stirring. Dialysis is thus feasible for oligonucleotides of degree of polymerisation n 2 6 (molecular weight 2 2000 dalton). Dialysis has been used for tetranucleotides ( n =4) but considerable losses are experienced (Vanderbussche and Fiers 1966; Lee and Deniset 1967). Dialysis tubing should be pre-treated by boiling in 0.1 M EDTA or bicarbonate solutions or both successively, and thoroughly rinsed in distilled water. The tubing and washing solutions should be kept scrupulously clean and free of nucleases. Gloves must be worn to avoid contamination with nucleases secreted through the skin. Washing the dialysis bags, particularly the inside, removes the plasticiser which would otherwise contaminate the oligonucleotides, and boiling reduces the pore size. The simplest dialysis procedure is to tie a double knot firmly in one end of treated dialysis tubing, add the solution to be dialysed and then tie another double knot above the solution. The bag is then placed in a large (usually 1-5 1) beaker of the solution to be dialysed against. The solution is stirred for at least 8 hr at 4°C. Some space for expansion should be left in the top of the tube because it is difficult to empty an overfilled bag and in extreme cases of expansion during dialysis the bag could burst. The time to reach equilibrium can be approximately halved by

Ch. 6

DESALTING A N D CONCENTRATING

299

using a rocking dialyser arranged so that an air bubble in the dialysis tube moves up and down the tube as the apparatus rocks so providing very good mixing. In principle the time could be reduced further by forcing the solution through an ultrafiltration membrane under pressure but we have no experience of the application of this method to oligonucleotides. Ultrafiltration could also be used for concentrating oligonucleotide solutions. Oligonucleotide solutions can be concentrated using ordinary dialysis tubing if a high concentration of a high molecular weight solute is introduced into the water outside the dialysis bag. Polyvinylpyrolidone (PVP) can be used, but Ficoll, a dextran polymer, sold by Pharmacia Fine Chemicals AB, gives generally cleaner solutions. The oligonucleotide solution is dialysed against 20% Ficoll in water. Care is necessary to see that the volume of oligonucleotide solution does not get so low that major losses are incurred in the knot of the dialysis tubing. Alternatively, the oligonucleotide solution can be placed in a measuring cylinder, and dry Ficoll in a dialysis bag lowered in so that the final volume of the oligonucleotide solution is below the end of the bag (Fig. 6.1).

6.6. Gel filtration Oligonucleotide solutions with molecular weight 2 1000 dalton can be desalted on Sephadex G-10 or Biogel P-2. Larger oligonucleotides (m.w. > 5000 dalton) are usually desalted on Sephadex G-25; other gels can also be used. The gels are examples of cross-linked dextran or poly-acrylamide gels respectively, formed into beads which act as molecular sieves. Their use in gel chromatography or molecular sieve chromatography or gel filtration, has been fully discussed in this series (Fischer 1969), so only an outline appears here and at other places in this manual ( 5 1.1). The gel beads (‘fine’ grade: 200-400 mesh) are pre-swollen by soaking in buffer overnight and then packed into a 3 H O cm long column whose volume is at least 3 times the volume of the sample Subjecr w d e i p. 487

300

SEPARATION METHODS FOR NUCLEIC AClDS

containing Ficoll

-measuring

cylinder

concentrat ion

magnetic stirrer

Fig. 6.1. Concentration of macromolecules by dialysis against Ficoll.

to be desalted. The columns are packed and operated at relatively low pressures, a few tenths of a metre of water, otherwise the column techniques are similar to those described previously. The column is eluted with 0.01 M ammonium carbonate or other very dilute volatile buffer at 10-20 ml cm-2 hr-'. It may be possible, particularly with polyacrylamide gels like Biogel P-2 to elute with water but secondary binding effects may occur (6 2.3.3). Such dilute buffer solutions can be easily removed by freeze drying or rotary evaporation. The oligonucleotide solution is loaded carefully on the top of the column. If it contains, for example, 7 M urea it can easily be layered on the top of the column under the eluting buffer. The top of the column can be protected by a filter paper. If there is no density difference the top of the column is allowed to run almost dry and then the solution carefully allowed to soak in followed by elution buffer. If a specially designed column end piece is used the sample

Ch. 6

DESALTING AND CONCENTRATING

30 1

can be introduced through it but no air bubbles must be introduced in the process. The column effluent is monitored and the oligonucleotide solution, which emerges first, is collected. These columns are very reproducible, in volume terms, and it is not necessary to monitor a column once it has been calibrated. When at least one column volume of eluting buffer has been fed into the column it may be used again immediately without further regeneration provided no secondary binding has taken place. Concentration can be achieved by gel filtration but this method does not appear to be widely used (Fischer 1969). Basically, dry gel is added to the dilute solution and in swelling the gel takes up solvent leaving high molecular weight solute outside the beads. The solution outside the beads is recovered by centrifugation in a basket centrifuge (from MSE Ltd.) or a swing out rotor fitted with filtration tubes. The centrifuge conditions are important.

CHAPTER 7

Essentials of gel electrophoresis

Zone electrophoresis in stabilised media has been used for many years for the fractionation of protein mixtures. Until the introduction of starch gels by Smithies in 1955, electrophoretic separations depended entirely on the existence of differences in effective charge between protein species. In general, the electrophoretic mobility is determined by the net electrostatic force acting on the particle, and this depends on the potential gradient, the effective charge (taking account of counter-ion binding, and shielded charges) and the frictional resistance. In nucleic acids, the constancy of the charge:mass ratio limits the possibility of separation by purely electrophoretic means, for each added unit of charge brings with it an additional unit of frictional resistance. This compensation is clearly shown by the experiments of Olivera et al. (1964), in which constant mobility in free zone electrophoresis is maintained over a wide range of molecular weights. It is true that this law may break down in extreme cases, and Shack and Bynum (1964) showed that native and denatured DNA could be separated in such a system. The reasons are two-fold: on the one hand the frictional properties of native and denatured DNA differ grossly, and on the other hand their electrostatic free energies are widely disparate, and are reflected in the extents of charge neutralisation by counter-ions. The possibilities of more general application of electrophoresis to the separation of nucleic acids arose with the advent of gels of controlled pore size for zone electrophoresis. Although agar gels were 302

Ch. 7

303

ESSENTIALS OF GEL ELECTROPHORESIS

first used for electrophoresis of proteins by Gordon and his associates in 1948, and retardation effects were shown by the variation of gel concentration, the use of ‘molecular sieving’ in electrophoresis was not really exploited until the advent of starch gels. Smithies (1955) showed that the electrophoretic mobility of proteins in high-concentration gels of partly hydrolysed potato starch was controlled by their molecular weights as much as by their charge. Polyacrylamide gels, containing covalent cross-links, were first used by Raymond and Weintraub (1959) and represented a considerable advance over starch, because stable gels could be prepared over a very much wider range of matrix concentration, because other aspects of the composition, such as the cross-linker concentration, could be controlled, and because charged impurities, which give rise to adsorption and electro-osmotic effects, could be almost wholly eliminated; moreover, the high transparency of the gels rendered them ideal for densitometry. The potential of polyacrylamide gel electrophoresis for separation of nucleic acids was not demonstrated until 1964, but since then the method has shown itself to possess unsurpassed resolution over an enormous range of molecular weights, and its possibilities are still far from exhausted. For analytical work it would be no exaggeration to claim that the method has largely superceded sucrose gradient centrifugaion. Simultaneously with the exploitation of polyacrylamide gels, as applied to RNA, other workers, primarily in Bulgaria, were developing the use of agar and agarose gels. These are less versatile than polyacrylamide, in that the practicable range of agar concentration is only about 0.54%, whereas polyacrylamide can be used between about 2.2 and 30%. However, in the high-molecular weight range, agar gels are just as effective, and the upper molecular-weight limit is higher. They have, as we shall see, some other advantages. Recently, ways have been developed of combining both media for the study of highmolecular weight materials. Starch gels were also shown (Beney and Szekely 1966) to permit good fractionation of RNA, but it now seems unlikely that they would offer any advantage over polyacrylamide, agar or agarose. Suhjrcr

I&\

p 487

304


7.1. General principles of gel electrophoresis In gel chromatography (using Sephadex, Sepharose or Bio-gel, which consist respectively of cross-linked dextran, agarose or cross-linked polyacrylamide), the solution of macromolecules is passed through a column of beads of the given material, and the solute molecules penetrate the solvated matrix to an extent which depends on their Stokes radii. The separation will depend then on a partition coefficient for the distribution of the solutes between the inside of the bead and the free solvent outside. The molecules which are least able to penetrate will be excluded from the beads and will emerge from the column without retardation. Smaller molecules will be progressively retarded by restricted diffusion within the beads and will emerge only at higher volumes of eluate (see Fischer, Vol. 1, Part 2 of this series). In gel electrophoresis, the solute molecules are impelled towards the electrode by a force equal to the product of the potential gradient and their charge. They will be retarded by the inherent hydrodynamic drag, that is to say the product of the velocity and the frictional coefficient, and in particular by the size-dependent probability that they will collide with a matrix molecule. The essentially similar natures of chromatographic gel filtration and gel electrophoretic fractionation has been noted (Rodbard and Chrambach 1970), and will be briefly discussed ($ 7.3). An important feature of gel electrophoresis is that diffusion is restricted in the matrix, and remarkably sharp zones are consequently obtained. This is discussed in a paper by Richards and Lecanidou ( 1971), which should be consulted for a quantitative treatment. The electrophoretic mobility of a species in agarose or polyacrylamide gel is determined by the Stokes radius, the effective charge, the nature of the counter-ions, the potential gradient, the concentration of agarose or polyacrylamide and, to a lesser extent, by the number of cross-links. For a solute molecule of appreciable size there will be a gel concentration at which the mobility is so diminished that the molecules do not enter the gel. The mobility in fact is a smooth function of the gel concentration, which must therefore

Ch. 7

305


be selected to suit the molecular-weight range of interest. The mechanical properties of the gel change with acrylamide concentration, and can be modified within limits by adjusting the relative number of cross-links. Most workers use a commercial 20: 1 mixture of acrylamide and methylene bisacrylamide, called 'Cyanogum 41,' which is satisfactory over a wide range, but not at very low or at very high concentration of polymer. One of the advantages, as we have noted, of transparent media, such as agar or polyacrylamide, is that they are suitable for densitometry. The densitometry of stained nucleic acid zones presents no difficulties, but the possibility also exists of direct recording of zones by scanning in the ultraviolet. This has considerable advantages for purposes of quantitation, but raises technical problems. Agar and agarose, being polysaccharides, are transparent in the ultraviolet, and the background at 260 nm in gels of good quality arises almost entirely from scattering. This largely vanishes when the gel is dried, and the ensuing transparent films are very suitable for scanning and for spectrophotometry in general. Polyacrylamide contains the amide chromophore, which absorbs below about 245 nm. There are clearly, however, other impurities of unknown character, for gels begin to absorb markedly below about 280 nm. The absorption can be diminishedby recrystallisation of the monomer before polymerisation, or by prolonged pre-electrophoresis. Scanning at 260 nm is then feasible, especially in gels of low concentration. Densitometry at 280 nm is satisfactory even in concentrated gels, and the loss in sensitivity relative to 260 nm is only a factor of about two. It will be clear that ultraviolet densitometry of nucleic acids is advantageous, in consequence of their high absorption. Moreover all RNA species at normal ionic strengths and temperatures have a similar molar absorptivity (in the range of 7200-7800 per phosphorus at 260 nm) so that the integrated absorbance at 260 nm will be proportional to concentration to within about 10%. One of the most felicitous features of electrophoresis of nucleic acids in gels is the dependence of the mobility on size. Consequently, if a calibration can be set up in terms of known species, gel electroPrhJ"'r

lIldC\

1' 487

306


phoresis can be used for the rapid determination of molecular weights of unknown samples. Clearly this cannot be carried too far, for if there are gross conformational differences - for example in comparing a fully helical rod-like species with one that is only partly base-paired the method will be no more applicable than similar empirical determinations based on parameters such as the sedimentation coefficient or viscosity. To nullify such limitations, methods must be used in which all species are reduced to true homologues, most feasibly by complete elimination of base-pairing.

7.2. The chemistry of gel formation Agar is a polysaccharide extracted from various red algae. The algae are boiled in water, acidified with acetic acid, and the extract is filtered and allowed to gel. An exudate containing impurities is removed on repeated freezing and thawing. The product is however impure, since it contains, in addition to the polymers agarose and agaropectin, various low molecular weight compounds and inorganic ions. On acetylation agarose acetate and agaropectin acetate are

Fig. 7.1. Structure of agarose L: a) repeating unit; b) section of polymer.

formed. These may be separated and the former is reconverted to agarose by saponification. The agarose polymer consists of repeating disaccharide units composed of fl-D-galactopyranose and 3,6-anhydro-L-galactose termed agarobiose (Fig. 7.1). Interchain hydrogen bonds are presumed to form the cross-links which lead to the formation of gels.

Ch. I

307


In contrast to agarose, polyacrylamide is a synthetic polymer, and gel formation depends on chemical cross-linking. The bifunctional acrylamide derivative, N,N’-methylene-bis-acrylamideis included in the polymerisation mixture to insert cross-links between chains of I I

F NH

I

co I

- CHz - CH - lCH2 - CHI, I

co

- CH2

- CH - CHZ I

co

- CHI,I

co

I

I

I

NH2

NH I

NHZ

NH

I

co I - CHz - CH - CH2 - CH - CH2 - CH - lCHz - CHInI

I

co

co

co

NH I

NH2

NHZ

I

I

Cl$

-

I

I

NH

I

co

- CHI Fig. 7.2. Structure of polyacrylamide cross linked by N,N’-methylene-bisacrylamide.

polyacrylamide (Fig. 7.2). However cross-links may also be introduced with other types of bifunctional reagent. Diacrylate or tartarimide cross-links (Fig. 8.1) are particularly useful because they are readily hydrolysed, and the gels can consequently be dissolved in alkali or acid respectively after electrophoresis for recovery of the RNA fractions, e.g. for radioactivity measurements. Subjeci index p. 487

308


The proportion of cross-linkingreagent may be varied to generate gels with different mechanical properties. The lower the cross-linker concentration the more deformable the gel. Polymerisation in the absence of cross-linking reagent produces a viscous solution of high molecular-weight polyacrylamide. A higher percentage of cross-linker is required in dilute acrylamide gels to provide mechanical strength. Conversely,concentrated gels require a low cross-linker concentration if they are not to be brittle and unmanageable. Gels with a relatively low number of cross-links swell to a greater extent on immersion in fixative solutions after electrophoresis. The percentage of crosslinker will alter the relative mobility of RNA molecules to some extent, in accordance with change in pore size (Fawcett and Morris 1966). Richards et al. (1965) give the following equation as a guide for the selection of cross-linker concentrations: 1OOR - 6.5 = 0.3C

(74

where C is the total acrylamide concentration and R is the fraction by weight of cross-linker in the monomer mixture. Cyanogum 41, the mixture of acrylamide and N,N’-methylene-bis-acrylamidein the proportions 95:5 by weight, is commercially available from a variety of suppliers, and can be used to form gels over a wide concentration range. It has been shown that the effective pore size is apparently minimised by this proportion of cross-linking agent (Fawcett and Morris 1966). Moreover the resulting gels have the highest optical clarity, because the scattering is also minimal. These effects will be further considered later (8 7.5.4). Polymerisation is initiated by free radicals which can be generated in a number of ways. The most commonly used reagents for this purpose are riboflavin and ammonium persulfate, or Fenton’s reagent for acidic gels down to pH 1 (Jordan and Raymond 1969). Free radicals are generated from riboflavin by irradiation with near-ultraviolet light, the most convenient source of which is an ordinary fluorescent strip-light. Small concentrations of amines such as tetramethyl ethylenediamine (TEMED) or dimethylaminoethyl cyanide (DMAEC) are included as catalysts. The rate at which the gel forms

Ch. 7


309

is proportional to the product of the concentrations of the free radical generator and the catalyst, and also depends on the temperature. The polymerisation is effectively inhibited by oxygen, so that the acrylamide solution is best degassed under suction, and contact of the surface with air must be prevented.

7.3. Theory of molecular sieving No comprehensivetheory of molecular sieving processes as yet exists. A number of semi-quantitativetreatments have however been evolved, and in general take as their starting point Ogston’s model, in which the gel is represented as a random arrangement of rods or strings of beads (Ogston 1958; Ogston and Phelps 1961); from the dimensions of these, the volume fraction available to a spherical molecule as it migrates through the matrix can be derived. The theory may be simply applied to gel filtration by assuming with Laurent and Killander( 1964) that Ogston’s fractional volume for a given molecular species, which is given by f

- ,-Id4

(74 (where I is the length of matrix fibre per unit volume, and s the surface area of the molecule) can be equated with partition coefficient f for distribution of that species between the matrix and the free solvent. Morris ( 1966) first suggested the analogous assumption for gel electrophoresis, that f could be equated with the ratio f o of the electrophoretic mobility ( u )in the gel to that at zero gel concentration, i.e. f =duo (7.3) In other words, the mobility of an assemblage of identical molecules through the gel is proportional to the volume fraction from which they are not excluded by the gel matrix material. The predictions according to this scheme for the dependence of mobility on gel concentration and Stokes radius are in good qualitative accord with observation for a series of native proteins (Morris 1966). This moreover provides a rational link between the mobility of a given V-

Suhiecr index p 487

310


molecular species in gel electrophoresis and its elution volume in gel filtration. This approach has been elaborated by Rodbard and Chrambach (1970). They consider the volumes available to the migrating molecule both in the Ogston model of a matrix of long thin rods, and in case of a matrix composed of randomly disposed planes. These two extreme cases (infinite rods and rods of zero length, i.e. points) lead to different expressions for J and consequently for the average pore size. The average pore radius is proportional respectively to T-* and T-’ in the two cases, where T is the total concentration of gel matrix in monomer units. These expectations are borne out for gels of very low and very high cross-linker concentrations, which correspond closely to the two ideal situations (Fawcett and Morris 1966). Further, Rodbard and Chrambach define a ‘retardation coefficient’, K,, such that log u/uo = K , T , (7.4) a relation that is known to hold good for a variety of molecules in various systems, the cross-linker concentration being held constant. The retardation coefficient is clearly related to the mean pore radius, and the effective radius, R, of the migrating molecule. In the two limiting cases, as before:

K i = c , ( R + r ) and K i = c,(R+r) (7.5) c1 and c2 being constants, and r the fibre radius. From experimental data on proteins Rodbard and Chrambach obtain the reasonable value of 7 8, for r, in remarkable agreement with 8 8, value derived from gel filtration on beads of the same material. With RNA, the effectivevalue of R, and its physical significance, is less easily defined and different values for r are obtained (Richards and Lecanidou 1971).

7.4. Theoretical aspects of factors affecting resolution This problem has been considered by a number of workers, but the most pertinent study, which moreover is directed towards electro-

Ch. 7


31 1

phoresis of RNA, comes from Richards and Lecanidou (1971). They consider in the first place the effect of diffusion, and make the assumption that the free-solution diffusion coefficient, Do, of a migrating molecule is attenuated to the same degree by the gel as is the free solution mobility, that is DIDo = u/uo =f

(7.6)

Since both these transport processes depend on the frictional coeficient of the molecule in the same way, this is a very reasonable assumption, and explains why under conditions of severe retardation zones remain extremely sharp. By application of Fick’s laws, Richards and Lecanidou determine the evolution of the zone half-width with time, and derive the important result, confirmed by experiment, that the zone-width, provided the ratio of diffusion coefficient to retardation factor is of the right order, depends to only a slight extent on the width of the starting zone applied at one end of the gel. For typical conditions there is no detectable loss in resolution for an initial zone width of up to 2 mm, and even for a 1 cm column of RNA solution applied to the gel there is only a 25% increase in band width. When the total mass of sample is increased beyond a critical value, conductivity disturbances within the zones lead to spreading and asymmetry. The point at which this effect becomes obtrusive will depend, in a manner that has been theoretically derived, on the ionic strength of the buffer and the mobility of its ions: Richards and Lecanidou suggest for a typical system, such as they use (see also below), a loading of 0.5 pg in cylindrical gels of 0.5 cm diameter in 0.05 M ionic strength buffer. At higher ionic strength higher loadings can be tolerated without loss of resolution. 7.4.1. Gel concentration for best resolution Rodbard and Chrambach (1970) evaluate T,,,, the gel concentration corresponding to maximisation of 6u/u, by differentiation,and obtain K, between l/e and l/e2, an expression which leads to the result that maximal separation is to be expected when the gel concentration, Subject index p . 487

312


Tmax, is 2 / ( K 1 + K 2 ) ,where K 1 and K 2 are the value of K , (see eqs. 7.4 and 7.5) for the two species to be separated. However, because of diffusion, the conditions for maximal separation (of the zone peaks) do not correspond to optimal resolution. Instead the optimal gel concentration, T,,,, is given by 4/(K, + K J . Moreover in the case of a set of homologues of identical free solution mobilities, the mobility at Tmxis about l/e, or 0.37 of the free solution mobility, and that at T,,, l/e2 or 0.135 of the same. This relationship I

I

1

I

I

-

-

XT

Fig. 7.3. Gel concentration for optimal separation (after Rodbard et al. 1974). The situation (mobilities of components A and B converging at zero concentration) is the ‘size isomerism’ case of Rodbard et al. T is the total acrylamide concentration. T,, is the gel concentration corresponding to maximum resolution, regardless of zone width, and that for optimal separation of the two components. p is the mobility of RNA.

zp,

Ch. I


313

in terms of ‘Ferguson plots’, is illustrated in Fig. 7.3 (Rodbard et al. 1974). It has to be recognised that there is thus far no full and satisfactory theory of gel electrophoresis and in general the best that can be expected of theoretical analyses is the rationalisation of observed effects. Their practical impact is therefore slight, and workers in the field have preferred to avail themselves of empirically derived rules in designing fractionation experiments. As to the choice of gel concentration, which is the most important experimental variable, Peacock and Dingman (1968)have suggested that optimum resolution is obtained when the mean molecular weight is approximately half the molecular weight corresponding to zero mobility, M,, obtained by extrapolation (Fig. 7.4). From this figure an optimum (or at least satisfactory) gel concentration can be chosen for determination of molecular weights of RNA in any given range. It can be seen that at high acrylamide concentrations the change in M , with gel concentration approaches a limiting value. Thus for oligonucleotides it should be advantageous to choose a gel of the highest workable concentration. Results on oligodeoxynucleotides(8 10.1) show the clear resolution of successive oligomers of (dAT), differing by only one base pair. Calculations by Richards and Lecanidou (1971) suggest that for larger species under optimal conditions comparable resolution may be obtained. The relation (Boedtker 1960) between s ; , , ~measured at sufficiently high ionic strength (say 0.1-0.2) and molecular weight for typical single stranded RNAs is S;O,~

= 0.02

MO.’

(7-7)

The more concentrated the polyacrylamide the more slowly the RNA molecules will migrate. Typical curves of mobility against gel concentration are shown in Fig. 7.Ab. Beyond a certain concentration the mobility will become so low that within any reasonable time the distance of migration is insufficientto permit satisfactory separation. Indeed, gradients of polyacrylamide concentration have been used, in which the electrophoretic mobility of each species diminishes with Subject index p. 487

314

SEPARATEON METHODS FOR NUCLEIC ACIDS

\ \

I I

\ \

\

\

\ \

\ \

\

I 1

\

\

b

\

\ \

\

\ \

\

\

\ \

1 I I I I I I I

distance of migration, until in principle it becomes essentially stationary at a position in the gradient depending only on its Stokes radius. This is referred to as pore-limit electrophoresis.Since diffusion is strongly restricted in concentrated gels, a relatively high gel concentration is advantageous for good resolution (see Richards and Lecanidou 1971). In practice the gel concentration is selected to be a compromise between resolution and time of experiment. Spectacular fractionations have been achieved in very long runs in polyacrylamide gels of high concentration, as will be seen later ( 5 9.2.2).

Ch. 7

315


1001

0.11

I

1 I

1

I

I

I

I

1

1

I

I 4 5 6 TOTAL GEL CONCENTRATION ( O h )

2

3

Fig. 7.4. a) Molecular weight-mobility relationships for electrophoresis of RNA in composite polyacrylamide-agarose gels of the stated acrylamide concentration (%) (adapted from Peacock and Dingman 1968). b) Dependence of mobility on gel concentration for single-stranded (dotted lines), and double-helical (rod-like) RNAs (full lines) (Fisher and Dingman 1971).

7.5. Experimental methods Gel electrophoresis is variously used to identify and characterise species in a mixture, to determine their relative concentrations, to measure their radioactivity, and to isolate them in modest bulk. Sg,hjerl index p. 4/37

316


Microelectrophoresis, for such purposes as the analysis of RNA from a single cell (say g), clearly provides a special challenge. We consider these aspects in turn. The same general principles apply to electrophoresis in tubes or rectangular slabs of polyacrylamide, agar or agarose or their mixtures. The most important manipulative factor determining the regularity of zones in the electrophoretic pattern is the evenness of the gel surface at the point at which the sample is applied. A flat meniscus prior to setting is obtained in tubes by layering the liquid acrylamide monomer solution with buffer. An even surface in flat gels is ensured by layering (in the case of vertical gels) or by using a plastic mould which forms the sample slots (either vertical or horizontal gels). 7.5.1. Apparatus for basic methods 7.5.1.1. Disc gels A common scheme for polyacrylamide gel electrophoresis in tubes is the ‘disc’ method. This was so named because of a sharpening procedure (Kohlrausch regulation), which was introduced by Ornstein (1964) and Davis (1964) with the object of causing the components to stack up in contiguous discs in a cylindrical gel of low concentration before entering the higher concentration gel, in which the separation actually occurs. This procedure, which is still often used, depends on the use of discontinuous buffer systems, in which an ion front from the reservoir buffer moves through the sample and causes macromolecules to migrate into an increasing potential gradient, and so accelerateand collect into a very thin zone. It is now recognised (Hjertkn et al. 1965) that sharpening of the sample zone can be achieved in other ways: applying the sample in a medium of low ionic strength compared with the gel buffer leads to an increased voltage drop across the volume element containing the sample which leads again to an acceleration within this volume, with consequent zone sharpening. In any case, however, there is always a discontinuity in mobility at the solution/gel interface because of retardation by the gel, and the sample therefore tends to accumulate in a narrow band at the top of the gel before entering. For this

qX174 R F &' TYMV

#XI74 I6srRNA

1051

1.0

20

3.0

I

4.0

Relative electrophoretic mobility (cm)

kr? rn

ELi

0

Fig. 7.5. Some typical molecular weight-mobility (or sedimentation coefficient) relationships for electrophoresis of RNA in polyacrylamide gels, showing conformational effects: a) set of single-stranded viral and ribosomal RNAs in 2.4% gel (Bishop et al. 1967). b) shows a typical separation of different species on a single gel, displayed as a densitometric trace; c) native single-stranded RNAs ( O ) ,formaldehyde treated RNAs (O), polyriboadenylic acid fractions with ( A ) and without ( A) formaldehyde treatment, and polyribouridylic acid fractions (0) treated with formaldehyde. These data show that the mobilities are little affected by single-stranded stacking, but substantially by partial base pairing and that formaldehyde treatment is inadequate to reduce RNA to the conformation of a single-stranded unpaired polynucleotide (from Pinder et al., 1974).

v,

: 4

318


reason the discontinuous system, with matching of ionic mobilities, has been discarded by most workers, and so especially has the use of 'spacer gel'. We do not believe that it offers significant advantages in routine use and its elimination certainly halves the labour involved in setting up an experiment (see however Gordon, this series Vol. 1 supplement where disc electrophoresis of proteins is discussed). However, the configuration of the apparatus devised by Ornstein and Davis has been retained in many laboratories, and it is available in several commercial versions. It is shown in Fig. 7.6. Glass tubes approximately 0.6 (i.d.) x 8 cm are convenient and are retained in the upper reservoir by grommets. The electrodes are generally of platinum wire. The reservoir volume should be no less than about 0.5 1, for otherwise serious contamination with electrode products in the time of an average run is hard to avoid. External cooling, say with an electric fan, is often desirable. This apparatus can be improved by arranging for the tubes of gel to be immersed in buffer for the greater part of their length, to provide a heat sink (cf. Fig. 7.7). A considerable number of tubes (often 6-10) can be accommodated in such an apparatus. It is moderately convenient in use, will take large sample volumes, and is almost foolproof. One of the disadvantages of the type of apparatus of circular design shown is the absence of any barrier to prevent the diffusion of electrode products through the buffer and into the gel. If this is allowed to happen, in the course of a long run, the pH will change, alien ions, including protons, may migrate into the gel and frequently react with the polyacrylamide matrix, and precipitate the RNA. This can be avoided by changing the buffer at intervals (when the pH has changed detectably to pH-paper), and using a good buffer at sufficientconcentration. An alternative solution is to use an apparatus of rectangular design, incorporating a barrier, such as a screen containing holes with porous plugs between an electrode chamber, and the bulk of the buffer compartments (Fig. 7.7). Alternatively, one may use external electrodes, e.g. reversible calomel electrodes, consisting of a paste of mercury and calomel under a saturated potassium chloride solution. Contact with the buffer compartments is best made

61f

SISWOHdOHL33’I3 ’I30 dO S’IVILN3SS3

1’

/

Fig. 7.7. Modified design of cylindrical gel analytical electrophoresis apparatus. One of the design features, which may also be embodied in the circular layout (Fig. 7.6) is that the tubes are immersed in the lower reservoir buffer for better heat dissipation. A further advantage in this system is the presence of a barrier which impedes the diffusion of electrode products into the buffer solution. 320

Ch. 7


321

by means of agar or polyacrylamide bridges containing potassium chloride solution, about 1 M. These survive for many runs, and do not leak much salt into the buffer. There is of course an appreciable voltage drop across the bridges so that a greater voltage will be required from the power supply than in the conventional arrangement. In one electrode, mercury is oxidised to calomel; in the other, calomel is reduced to mercury. Consequently if the electrode connections are interchanged between successive runs, the electrodes will maintain themselvesindefinitely, so long as the platinum wires remain immersed in the paste at the bottom (Fig. 7.8).

Fig. 7.8. Reversible mercury-calomel electrode with the parts consisting of a 250 ml widemouthed reagent bottle, a vented and drilled rubber stopper, a soft glass tube with platinum wire sealed into the end, mercury, a wire to the output of the power supply, saturated KCl, and a glass tube containing 5% polyacrylamide gel made up in 1 M KCI. The tube dips into the buffer reservoir of the electrophoresis apparatus. The anode and cathode should be interchanged after each run. This arrangement will prevent all contamination of the buffer with electrode products even in very long runs (J. C. Pinder, Ph.D. Thesis, London University, 1974). Subject index p. 487

322

SEPARATION METHODS FOR NUCLBIC ACIDS

It requires a measure of skill and practice to achieve an equal voltage drop in all tubes. The tubes themselves should be of high quality to ensure an even bore, and must all be filled to precisely the same level, using scratch marks on the glass. The apparatus must be such that the tubes are symmetrically distributed around the electrode or bridge, and blank gels should be used rather than blocking the holes for tubes if it is not required to run the full complement of gels (unless of course a symmetrical arrangement can be maintained). Even so, comparison of zones by dead-reckoning is difficult (though not impossible: see Lewicki and Sinskey, 1970), and the split-gel technique (Clarke, 1964; Leboy et al. 1964), where two samples are applied to a single tube, usually with a dividing strip, extending, say, 1 cm down from the top, may be useful.

Materials Cylindrical gels of acrylamide may be cast in glass or Perspex (Lucite, Plexiglass) tubes; gels containing agarose do not adhere and tend to slide out, so tubes cannot be used. Gels adhere less readily to Perspex, which is consequently often recommended for concentrated gels. The replacement of gels into Perspex tubes (which can be desirable when it is required to equilibrate the gel with a buffer different from that used for the setting reaction) is easier. However very dilute (eg. 2.5% or below) gels tend to slip even out of glass tubes (and more especially Perspex), and may have to be supported by a rim of plastic tubing the external diameter of which is such that it fits tightly in the tube at the bottom, or by wrapping a layer of surgical gauze around the bottom of the tube, to which it can be secured by a rubber band. Very concentrated gels present their own problems, and in particular are difficult to remove from the tubes without breaking. A disadvantage of Perspex is inferior heat dissipation. Gel dimensions The length and diameter of the tubes can be varied within limits. Tubes 6-10 cm long and 5-7 mm internal diameter are in common use. The diameter of the gel is limited by the need for adequate heat dissipation. The commonly observed occurrence of

Ch. I


323

curved zones may usually be attributed to heating effects; molecules migrate faster in the centre of the gel, where temperature is higher, than near the circumference. Richards and Lecanidou (1971) have demonstrated these effects under controlled conditions. Efficient cooling or the use of lower currents permits the use of thicker gels which may be useful for running relatively large samples. If the diameter of the gels is small, increased difficulty may be experienced in removing them from the tubes, especially at high acrylamide concentrations. Longer gels may be useful for increasing the distance between the separated zones by running for longer times at a given voltage drop per unit length of gel. The increased resolution so obtained may be particularly desirable for excising single zones for preparative purposes or for radioactive counting. Gels of up to 40 cm length have proved of value for preparative separations. At the same time it must be borne in mind that the longer the run, the greater the diffusional broadening of the zones. It is advantageous to use gel slabs (see below) rather than tubes for large migration distances because of the intractable nature of long cylindrical gels. 7.5.1.2. Flat gel beds: general considerations For direct comparison of samples, and many other convenient features, including better densitometric analysis, flat gels with samples applied in slots, are to be preferred. With agar and agarose the tubes do not, in any case work, since these gels do not adhere satisfactorily to the glass and tend to slide out. Gel slabs may be horizontal or vertical. Horizontal methods are alleged to suffer from electrodecantation effects: sample molecules in the application slot concentrate at the boundary of the gel, and the concentrated solution, being denser than the solvent, falls to the bottom of the slot. The result is an uneven distribution from top to bottom of the sample area, which leads to some loss of definition, since at all but very low concentration, the mobility is somewhat affected by concentration. Be this as it may, horizontal gels have been used in many laboratories, with excellent success. They lend themselves well moreover to the use of an efficient cooling system: e.g. the gel can be placed over a Suhjeci index p . 487

324


hollow metal plate, through which water is circulated (Vasu 1969). The reader’s attention may be drawn at this point to the application illustrated in Fig. 10.8. Here, despite theoretical reasons for preferring apparatus of vertical design, in practice equally good resolution may be obtained with horizontal gels which may be more convenient to handle. Perhaps the most rational geometry for polyacrylamide electrophoresis is that of vertical flat-bed systems (Akroyd 1968; De Wachter and Fiers 1972). The system was devised to allow the reproducible production of acrylamide slabs containing an acrylamide concentration gradient, but is altogether suitable for conventional use. Forms of all the systems mentioned here have been used in the authors’ laboratory and give good results. Vertical, flat beds, however, offer nearly all the advantages of the disc system with none of its limitations, and is particularly recommended. Few issues of the analytical journals go by without the appearance of a new design for gel electrophoresis apparatus, and there are no doubt many satisfactory alternatives to the three versions that we have discussed. Microelectrophoresisand preparative electrophoresis pose special problems, and will be considered separately (4 7.5.2.2; 4 7.5.2.3; ch. 9). 7.5.1.3. Horizontal flat gels A convenient apparatus for horizontal electrophoresis can be made from Perspex sheet and a glass plate as shown in Fig. 7.9, designed by Vasu (1969). The gel is poured into the space between the plastic mould carrying the sample slots and the bottom glass plate separated by rectangular Perspex spacers and closed off. After polymerisation, .the gel is inverted onto the glass plate which serves as a support during electrophoresis and the plastic mould is peeled away carefully, so as not to damage the sample slots. The gel can be soaked in buffer to remove contaminating chemicals or to introduce a different solvent. For electrophoresis, connections are made to reservoir buffer

Ch. 7


325

Fig. 7.9. Flat-gel horizontal electrophoresis apparatus. Design of Vasu (1969), and reproduced here with details supplied by him. The apparatus is made of Perspex and the base plate (1) is 1 8 . 6 ~13.0x0.5 cm. The vertical pieces (2, 3, 4, 5) are all 3 x 1 x 0.2 mm, and the sides are defined by slats, 12 and 13, respectively 10.8 x 0.5 x 0.3 cm and 11.2 x 0.5 x 0.3 cm, and 14 and 16, which are 20 x 1 x 0.2 cm. The glass upper plate, 15, is 1 6 x 11 xO.2 cm. This is placed on top of the base plate, resting on the slats (except the side piece 16) mounted as shown, 14 supported by 4 and 5, which are sealed to the plate. Likewise at the ends, are slats 12 and 13, the former supported by 2 and 3. The slot formers stick to the lower plate, and measure 1.2 x 0.2 x 0.2 cm. The acrylamide solution is poured down an inclined glass plate into the gap, and slat 16 is put in place. After polymerisation, the side slats are eased out with a razor blade and the device is inverted. The end pieces are likewise removed, and sheets of Whatman 6 MM paper are used to make contact with the reservoir, one layer above, another below the gel. After loading, and during the run, the gel is covered with a thin plastic sheet. Subject index p. 487

326


compartments with filter paper or plastic sponge wicks, cut to the width of the gels, and impregnated with buffer. Alternatively, one may cast a longer gel, which overlaps the supporting plate at both ends, and dips directly into the buffer compartments. During electrophoresis evaporation is prevented by covering the gel with polyethylene film or Saran, which is applied after the run has commenced and the sample has entered the gel. Alternatively (or additionally), the gel is enclosed in a humidity chamber. The Shandon Universal Electrophoresis tank is convenient for this purpose, though for long runs larger buffer compartments are desirable. The covering prevents drops of condensation which form on the lid of the apparatus from dropping onto the gel. In the absence of a humidity chamber it is probably best to spread a layer of paraffin oil on the surface of the plastic sheet in contact with the gel. If the plate carrying the gel is placed on a hollow brass block with an inlet and outlet for circulating tap water, good heat transfer is achieved and shorter runs at higher voltages are possible. 7.5.1.4. Vertical flat gels The simple apparatus for vertical flat bed electrophoresis (Fig. 7.10), described by Akroyd (1968), consists of two thin glass plates taped together and separated by spacers, which also form the sides of the compartment, and to which the glass plates are clamped. The gel sits between the glass plates and this sandwich is mounted vertically with the aid of plastic gaskets in a reservoir of ample volume. The cell stands in the lower reservoir, while the upper reservoir is placed behind the top of the cell and is connected to it by a filter paper wick. The height of the buffer over the gel should be the same as that in the buffer reservoirs to prevent siphoning.The electrodes are in the buffer compartments, and excessive trouble from contamination with electrolysisproducts is avoided by circulatingbuffer continuously between the upper and lower reservoir with a small pump (e.g. an aquarium pump). Alternatively, the buffer reservoir may contain bafiles to prevent electrode products from diffusing into the gel. A variant, also described by Akroyd ( 1968) employs water cooling which

Ch. 7

ESSENTIALS OF GEL ELECTTOPHORESIS

327

permits the current to be doubled. The home-made apparatus is as satisfactory as many which are commercially available. It can be modified to meet individual requirements; for example, 40 cm long glass plates have been used to increase the distance of the separations for the isolation of nucleic acids for sequence determination.

Fig. 7.10. Flat-gel vertical electrophoresis apparatus of the Akroyd type, showing three gels being run in parallel (De Wachter and Fiers 1971).

For filling, Akroyd supports his apparatus in a block of plasticine, to close the end of the gel compartment, and fills to within a few cm of the top of the glass and then overlayers with water. If tall gels are required and, because of the hydrostatic pressure, or for any other reason, there is leakage at the bottom, the gel can be set in two layers, a shallow layer at the bottom and, after this has set and the water layer been removed, the remainder of the gel solution is poured in. Sample compartments are made by thrusting tightly fitting pieces of rubber cut from rubber tubing between the plates. Using the same apparatus, a number of workers have preferred to construct a sample slot former from Perspex sheet which fits into the top of the cell, and around which the gel may set. This has the appearance of a wide-toothed comb and may be designed to permit one or several Subject indrr p, 487

328


Fig. 7.11. Flat-gel vertical electrophoresisapparatus: a more complex design, adaptable for the utilisation of several flat gels simultaneously. A is the anode, B an inlet for circulation of buffer in the upper reservoir, C, D the sample slot former, E the electrode, F the gel, H a silicone rubber gasket, J the lower reservoir and K the outlet for buffer circulation from the lower reservoir (by courtesy of Universal Scientific Co.).

Ch. 7


329

samples to be run. The width of the teeth from front to back may be slightly less than the width of the gel, thus presenting barriers to migration along the glass faces. For preparative purposes a single broad band of sample can be applied. A number of manufacturers have marketed apparatuses for vertical starch gel electrophoresis, which can equally well be used for polyacrylamide gels. A good design is that of Fig. 7.11. A commercial apparatus with water-cooled plates manufactured according to the design of Raymond (1964) by E-C Apparatus Corporation (Fig. 7.12) has been successfully used, notably by Peacock and Dingman (1968) for the separation of high molecular weight RNA in gels containing down to 3.5% acrylamide, or less, when agarose is included. An apparatus allowing multiple flat gels to be run in parallel is made by Universal Scientific Co.

Fig. 7.12. Flat-gel vertical electrophoresis apparatus after Raymond (1962), and manufactured by E-C Apparatus Co. The gel is supported on a sponge strip, F, and is confined between two cooling plates A and B. The samples are applied in slots, E, C and D are electrodes. G the reservoir buffer levels, and H a constant-level tube. Suhjrc I Index p. 487

330


7.5.2. Elaborations of basic methods 7.5.2.1. Two-dimensional gel electrophoresis

Two-dimensional gel electrophoresis of RNA has been used for two purposes. One general problem has been to identify RNA species migrating in the form of nucleoprotein particles, e.g. ribosomes and viruses. In this case, the buffer used in the second dimension is one in which the RNA and protein components are dissociated. RNA species of known size can be run in parallel with the sample in the second dimension. The strategy is simply to combine two analytical separations. A sample may first be run in a cylindrical gel, which is either sliced longitudinally to generate a flat strip or applied whole to a slab. It must then be embedded in new gel at the surface of the preformed slab to provide good contact. Again, the first dimension may be run in a flat gel, and a strip cut off for application to the second gel, which may be horizontal or vertical. In general it is advantageous to make the acrylamide concentration in the second dimension higher than in the first to bring about zone sharpening at the interface before the second separation begins. The second dimension may serve simply to improve separations. Enhanced resolution may be necessary because there is a wide range in the size of RNA species present in a mixture, in which case a low polyacrylamide gel concentration separates the high molecular weight RNA species and may be followed by electrophoresis at a higher concentration gel to optimise separation of the low molecular weight species (the magnitude of this type of effect is illustrated in Fig. 7.13). Alternatively, for the same purpose, a gel gradient may be used. The mobility of RNA species may be differently affected by the use of a denaturing solvent in the second dimension. This may occur by reason of the loss of secondary structure per se, or because of the dissociation of two or more strands in a base-paired complex in an enzymically 'nicked' molecule. In principle, compositional differences may be used to separate nucleic acid species of the same size and conformation, and therefore mobility at p H 7, by performing, in the

Ch. I


33 1

second dimension, zone electrophoresis at a pH in the ionisation range of the bases. Excellent results have been obtained using a neutral gel in one dimension, and either acid pH, or 6 M urea (or both) in the other (see Fig. 7.14)(De Wachter and Fiers 1972). 7.5.2.2. Microelectrophoresis The methods which have been discussed allow considerable scope for reduction of scale. There is now a sizeable literature of research on proteins involving electrophoretic analysis of the contents of single cells. The earliest successful results were obtained with fibres pulled from polyacrylamide gel, and subsequently two good procedures have been evolved, one based on cylindrical gels in glass capillaries, the other on very thin leaflets of gel (or microslabs). The capillary technique is probably the more generally tractable, and for protein work has been adapted for acrylamide gradients and for isoelectric focusing,as well as molecular weight determination by electrophoresis in sodium dodecyl sulfate. It has been successfully applied to the fractionation of RNA. The first application was that of Ringborg et al. (1968) and Egyhazi et al. (1969), who used composite agarosepolyacrylamide gels to fractionate the RNA of salivary glands of Chironomus larvae. The acrylamide solution is introduced by capillarity into 5 pl glass capillaries. The samples are loaded with a hand-pulled microcapillary, and in most respects the technique is similar to the conventional procedure already described. Naturally the manipulative techniques are much more exacting, and ingenious methods have been devised for loading known volumes of sample in the microlitre range, for microdensitometry of stained gels and for isolating components from a mixture by electrophoretic extraction. The diameter of the capillary is about 0.5 mm and separations may cover, say, 3 cm. Samples down to the nanogram level can be analysed. Results of remarkable quality have been obtained. We do not propose to deal here with the manipulative details of this specialisedaspect of gel electrophoresis,since they have been described in full detail by one of the innovators in the field, in a book devoted to a considerable extent to microelectrophoresis (Neuhoff 1973), and Subject index p 4 N 7 to which reference should be made.

332 SEPARATION METHODS FOR NUCLEIC ACIDS

Ch. 7 ESSENTIALS OF GEL ELECTROPHORESIS

Q

-

p

P v

Fig. 7.13. A) Autoradiograph of 32P-labelledtRNA mixture on flat gels, of 10, 5 and 20% acrylamide. B) Re-electrophoresis of 3 selected slices of these gels. The fist two slices were run at 10% in the second gel. The thud slice at M%(Ikemura and Dahlberg 1973). 333

Suhjerr index p. 487

v.

w

W

3 34 SEPARATION METmDS FOR NUCLEIC ACIDS

Ch. I


335

Fig. 7.14. a) Autoradiograph of two-dimensional gel showing resolution of lowmolecular weight RNA species. Note that zones on a common ordinate would be unresolved in the first dimension. The rectangle shows the position of embedding of the first-dimension gel. b) Map showing the dependence of migration rates in the two-dimensional system on base-composition. The effect of base composition depends on the fraction of bases present that have a pK in the acid range, which pertains to the second dimension (De Wachter and Fiers 1972). S u b p r m l r \ p 4x7

336


7.5.2.3. Preparative techniques Fairly wide use has been made of preparative gel electrophoresis in protein chemistry, and in principle there is no reason why the same procedures should not be adopted for use with nucleic acids which have the advantage that much may be accomplished with very small quantities of purified material. Thus, it is relatively easy in many situations to introduce radioactive label at very high levels and specific activity, and the use of 32Pfor this purpose offers a degree of sensitivity that cannot be matched in work on proteins. The extinction coefficients of nucleic acids are also very high in the ultraviolet, so that with say 20 pg in 1 ml or less it is possible to measure optical properties, thermal melting profiles, sedimentation coefficients, and even molecular weights by sedimentation equilibrium in an instrument equipped with scanner optics. Consequently, the sacrifice of resolution that, by a malign law of nature, always accompanies any attempt to scale up an analytical fractionation method is often at least partly avoided. There are two general strategies open to one: a) after electrophoresis, zones may be individually excised and the nucleic acid extracted, i) by diffusion into a buffer in which the slices are bathed, or by macerating or pulverising the gel by hand or in a commercial apparatus; ii) by using a gel with hydrolysable cross-links, or iii) by causing the zone to migrate out of the gel and into a confined volume of buffer by electrophoresis. The alternative approach is: b) direct electrophoretic elution from the bottom of a column of gel. The first approach (a) is used in sequencing work, usually allied to extraction by course (i): method (ii) is not generally satisfactory if the nucleic acid has to be recovered undamaged, for the hydrolysis conditions are rigorous. Before the zones can be extracted they must be located. This may be done by staining, although here again covalent damage is hard to avoid with any certainty; alternatively, an identical gel (or strip of the sample pattern from a flat gel) may be stained and matched against the unstained gel; or if the sample is radioactive, a rapid autoradiogram may often be obtained, which will locate the zones. Alternatively, the gel may be sliced into equal

Ch. 7


331

pieces along its entire length, each one extracted and the nucleic acid determined by ultraviolet absorption or radioactivity. A distancenucleic acid concentration profile can then be constructed. This of course is relatively arduous. Direct elution techniques involve the collection of the eluate by a continuous stream of buffer, which is then fed to a fraction collector, by way of an ultraviolet monitor if desired. Some examples of successful applications will be given in ch. 9. In principle this type of technique can be scaled up considerably, to the milligram level, but great care in the design of the apparatus is required if the separations are not to be vitiated by heating in the thick gels that are used, and if there is not to be distortion of the zones. Some loss of resolution, as we have remarked, appears inescapable. 7.5.3. Detection and ~ u a n t i t a rof i ~separated ~ components The simplest procedure for detection of RNA is to scan the gel, directly after the run, in the ultraviolet. Using recrystallised acrylamide, gels or subjecting the gel to pre-electrophoresis before sample application, scanning at 260 nm is possible. For gels of high concentration it will be necessary to operate at longer wavelength, with considerable loss of sensitivity. With agar and agarose there are no such problems, and scanning of the wet gel is possible, though prior drying gives the best results. If for one reason or another direct densitometry does not appear possible or desirable, then the gels may be stained. A further painless method of detection is due to Eisinger (1971): slab gels are cast between two plates, one of quartz and one of a fluorescent glass, such as uranium glass. When the gel is viewed under illumination from an ultraviolet lamp, the fluorescent glass gives rise to a yellow-green background on which the nucleic acid zones stand out as black bands. The progress of a run can be monitored in this way, and photographically recorded. 7.5.3.1. Staining Staining has a number of obvious advantages: in the first place it Suhlecr ,,trier 12 487

338


produces a permanent visual record of the experiment. Secondly, the eye is always better able to detect partly resolved zones than the densitometer. There are a number of reasons for this, one of which is the ‘stepped wedge’ illusion, whereby an inflexion in an intensity profile is perceived as a minimum (Holiday 1937). Thirdly, a densitometer cannot cope with the same enormous range of absorbance (optical densities) that the eye can discriminate, so that very weak or very strong bands will not be faithfully recorded relative to the remainder. Fourthly, zones in visible light can be recorded without interference from ultraviolet background absorption, and very much less obtrusion from scattering. Finally, many dyes can apparently be taken up in a ratio of one dye molecule per nucleotide (Stone and Bradley 1961), and since the molar residue absorptivity of an RNA is of the order of 7500 and that of a typical acridine dye is 50,000 or more, a gain in sensitivity of up to about ten can in principle result. On the other hand, using the familiar ‘stacking’ dyes, considerable deviations from Beer’s law would not be surprising (Steiner and Beers 1960; Stone and Bradley 1961), and not all RNA species will necessarily give rise to the same colour values in consequence of differences in stacking coefficients. Claims in the literature of linear response must be treated with caution: many possibilities exist of compensating errors, and, if precise quantitation is important, it is necessary to ensure that under the experimental conditions a linear law is indeed obeyed. The choice of stains is wide: acridine orange, pyronine Y,toluidine blue, methylene blue, gallocyanine, and ‘Stains-all’ (a carbocyanine dye, described in 5 8.3.1.3) have all been used. Acridine orange gives a somewhat lower sensitivity - though a large gain is possible by making use of its fluorescence - but at the same time it binds very strongly and is least easily removed by over-vigorous electrophoretic destaining. With agar gels, staining is best done after drying, using the same dye solution. Although it has been shown that diffusion of RNA in gels is a relatively slow process (Richards and Lecanidou 197l), it is considered highly advisable to precipitate the RNA when staining. Experience

Ch. I


339

now shows that if the stain is made up in a sufficiently acidic solution RNA will precipitate, and probably nothing is to be gained by including heavy metals in the staining solution. Destaining may be accomplished simply by soaking in the staining solvent, preferably in a dish attached to a rocking mechanism. For the impatient, electrophoretic destaining is recommended. Many apparatuses have been devised, and in general the dye is caused to migrate out across the shortest dimension of the gel. The tubes or slabs may be confined between porous plastic plates (Richards and Gratzer 1968) (Fig. 7.15 Shandon Scientific Co.), placed on a grid over charcoal (Stanton 1965) or simply contained in a plastic frame g

f

adbdc a

f

g

e

Fig. 7.1 5. Electrophoretic destaining apparatus (Richards and Gratzer 1968): cross sectional diagram ofelectrolytic destainer. Three gels are shown in the gel compartment (b) composed of a perspex (lucite) frame bounded by two sheets of Vyon porous plastic (d). This compartment rests in a holder made of two further perspex frames (a) joined along their bottom and sides. The gel compartment and the porous plastic are held in the holder by a fourth removable perspex frame (c). The holder fits tightly into a perspex tank (e) and thus separates the two buffer compartments (g), which are normally filled with 10% acetic acid. The potential is applied through two in carbon rods (0which are mounted in the tank through rubber grommets.

with electrodes of large area on either side (Fig. 7.16). For flat gels a good apparatus is available from Gradipore, in which 5% acetic acid is circulated by means of a pump, through cotton pellets which absorb the dye and are replaced when saturated. The danger in electrophoretic destaining is that the dye will be removed from the RNA zones as well as the background, and it is therefore important not to use too high a voltage. Patience at this stage is advised. Sub1er.r index p. 487

340

SEPARATION METHODS FOR NUCLEIC ACIDS I

I

I

I

Fig. 7.16. An alternative design for a destainer. The cylindrical holes contain the gels, the sides of which are left exposed. This frame is placed in a vessel containing destaining solution, with an electrode on either side (B. Tiplady, personal communication).

Model 2410 Linear tranmort

wrem

Fig. 7.17. The Gilford gel scanning attachment for spectrophotometer. The gel (cylindrical) is seated horizontally in a rectangular silica tray, which is driven across the beam in the linear transport assembly (by courtesy Gilford Instrument Co.).

Ch. 7


34 1

7.5.3.2. Densitometry Many densitometers are available, the best ones for the present purpose operating on monochromatic light. Instruments, such as the many designed for scanning of X-ray diffraction patterns or spectrographic plates (best known amongst which is the Joyce-Loebl microdensitometer), use white light, or at best a broad-band glass filter, and these must in principle give rise to erroneous results, for the unabsorbed part of the light, outside the absorption band of the dye, will appear as an apparently too high transmittance of the sample. Ideally in such an instrument, which operates by null-balance with an optical wedge, a wedge with the same spectral characteristics as the zones should be used, and can in fact be constructed to a good approximation. Such problems are obviated, however, if an instrument based on a monochromator is used, such as the Gilford 222 photometer fitted with the Model 2410-S Linear Transport System and 241 1 Adaptor (Fig. 7.17), or the Unicam SP1800 doublebeam spectrophotometer with an excellent scanning accessory. There are a number of other desirable characteristics of a densitometer for gels: a deep focus helps to sample the whole depth of a zone; the gel should be fairly near the detector to minimise scattering losses; and the stray light characteristics of the monochromator should be good enough to make it possible to cover a wide range of absorbances. Clearly the width of the slit image at the gel should be narrow enough to resolve close-lying zones. The zones as recorded (if on an absorbance, rather than a transmittance basis) can be quantitated by measuring the area under the peaks by planimetry, triangulation, cutting out and weighing or simply counting squares; or the densitometer may have an integrating facility. Many home-made and other commercial instruments have been constructed. Excellent results in the ultraviolet and visible have been obtained with an instrument designed and described by Tsanev and Staynov (1964) and Staynov and Stainov (1969). The gel may be laid on a silica plate (or glass for visible scanning), or it may, as in the Gilford design, be confined in a narrow Subject index p. 487

342


silica cell with flat walls. This is narrower than the gel which is compressed when introduced, and so does not present a curved gel surface to the beam. This is advantageous in that it minimises both adventitious lens effects due to curvature, and errors due to surface aberrations. In favourable conditions, and with a good monochromator, linearity of optical response up to an absorbance of 3 can be achieved. A number of cheaper densitometers utilising filters in place of monochromatic light are available. The Joyce-Loebl ‘Chromoscan’has been fairly widely used. This and the larger JoyceLoebl microdensitometer employ optical wedges to balance the sample absorbance. A glass filter can (and should) also be inserted so as to limit the wavelength-sensitivity. The use of an optical wedge with the absorption spectrum of the RNA-dye complex in such an instrument has been alluded to (Staynov, personal communication, has constructed such wedges by making up a dye-RNA solution with agarose and letting the gel set in a wedge-shaped container. The gel is then dried down on a glass plate). For any given densitometer and experimental circumstances, where the quantitation of the relative amounts of RNA (by staining intensity) is required, it should be stressed that it is prudent to check the linearity of the measured absorbance with RNA concentration, using known concentration increments of nucleic acid. The procedures for reducing the ultraviolet background of polyacrylamide gels to increase sensitivity for ultraviolet measurements are dealt with in § 8.2.1. 7.5.3.3. Recording and storage Instead of densitometric evaluation gels may readily be photographed by transmitted light. The simplest method is to make contact photographs, by placing the gel on a suitable photographic plate, and exposing it to light. This is not however always satisfactory, especially for cylindrical gels, and more versatile procedures are described in the experimental section ($ 8.3.1.4). The gels can be stored indefinitely in dilute acetic acid; cylindrical gels can be sealed into Perspex boxes of dimensions just sufficient to hold all the gels from a run, using chloroform, or a solution of

Ch. I


343

Perspex in chloroform to make the boxes watertight. Another convenient storage procedure is to put the gels in their dilute acetic acid wash into small flat-bottomed tubes (e.g. sample storage sets made by R. P. Cargille Laboratories, Inc.), and store them in the accompanying boxes which have an indexing list. Alternatively, the gels can be allowed to dry in air, when they become small, hard and transparent, the size of matchsticks. The zones are visible in such dried gels, but the gels can also be re-hydrated by immersing in water. An advantage of agar or agarose gels is that after drying on a glass plate they can be detached as a tough pliable film, ideal for densitometry, photography and storage. Unstained gels can be illuminated by an ultraviolet lamp (Woods glass filter) and if placed over a sheet of white paper the RNA components become clearly visible as dark zones on the yellow fluorescent background from the paper. The RNA does not diffuse during drying, and is suitable for all manner of spectroscopic studies. 7.5.3.4. Radioactivity Where radioactive components are being fractionated, there are two ways of recording the radioactivity of separated components in a gel. These are autoradiography and extraction followed by counting. The label may be tritium, carbon or phosphorus isotopes. The direct counting method is in many respects preferable to autoradiography. It is more suitable for tritium, is generally quicker to give results (except for a very high level of radioactivity), and is quantitative. A prerequisite is an efficient and accurate means of slicing the gels. Several good designs of gel slicers exist. The gel may either be sectioned with a row of blades or wires all at once, or it may be dealt with by a device of the bacon-slicer principle. Alternatively the gel macerator can be used in which the acrylamide is forced through a small orifice and pulped as it goes. Before transfer to the scintillation vials the RNA can be extracted by diffusion, alternatively the gel may bp dissolved by chemical brute force, or more gently if a hydrolysablecross-linkingreagent is used in its preparation. For plaque counting, the gel slices can be simply dried onto paper. Subject index p . 487

344


Autoradiography is rapid and simple with 32P, less so with the other available radioisotopes. Highly efficient techniques have been developed for use in sequencing, though these always demand strong labelling with 32P. Flat gels give best results in autoradiography, but cylindrical gels can be sliced longitudinally to give flat strips. Practical details for the exploitation of a selection of the more widely interesting methods are given in $ 8.3.1.5. 7.5.4. Gel composition As already noted (8 7.4) there are some empirical and semi-empirical guidelines to the choice of gel concentrations and composition. The concentration of bisacrylamide giving mechanically tractable gels is given by Eq. 7.1 ($ 7.8), though a fixed concentration of 5% of the total acylamide is used by many workers, as in the mixture Cyanogum 41. Gels of very high concentration are stiff and very difficult to remove from the tubes, and fragment easily. The practical higher limit is probably 25-30%. At the opposite extreme, gels of less than about 3% acrylamide are soft, and almost fluid below, perhaps, 2.2%. To obtain a gel of equivalent pore size which is more manageable, one may choose a somewhat lower concentration of acrylamide and raise the proportion of bisacrylamide, within limits (but see below). Peacock and Dingman (1968) introduced the addition of agarose (say 0.5%) as a mechanical stiffener where it is desired to operate at still lower acrylamide concentration. It may then of course be asked what advantage remains over agarose alone, which, as we have observed, gives excellent results in the high-molecular weight range, considering especially the greater effort involved in setting up an experiment (since the agarose has to be dissolved separately and maintained at a narrowly defined temperature for mixing with the acrylamide and other components). The chief advantage does appear to be great flexibility in the selection of effective pore size. Moreover at, say, 3% acrylamide which has a smaller pore size than agarose gels, the addition of agarose is still helpful if the gel is to be sectioned or sliced since these operations are very difficult with soft gels.

Ch. I

345

ESSENTlALS OF GEL ELECTROPHORESIS

The effect of acrylamide concentration on the gel characteristics must be considered, and is capable of theoretical analysis. Richards and Temple (1971) have considered in terms of the theory of rubbery elasticity the concentration-dependent properties of polyacrylamide gels. They introduce the concept of an ideal gel, defined by the relative values of the total monomer concentration To,and that of the cross-linker C. When T > To, where To is the value of T for an ideal gel, one has a 'crumpled gel', in which there is too much acrylamide, so that the polymer molecules are less extended than they would be in the unperturbed free solution state. Conversely, when T < To, there is not enough acrylamide polymer to link the ends of the bis molecules to polyacrylamide neighbours in the solution. This is a 'clustered gel', and if the polyacrylamide strands are not to be stretched beyond their normal solution end-to-end distance into extended chains, rather than coils, there must be clustering of molecules. A crumpled gel if soaked in water will reach equilibrium ( T = To),corresponding to the natural extension of polyacrylamide chains, by swelling, as is commonly observed in practice, when cylindrical gels are removed from the tube and soaked in staining and clearing solutions. Clustered gels will not swell and will have the characteristic property of turbidity, brought about by a non-uniform distribution of matrix molecules (clustering). This too has been often remarked on, in relation to gels of low acrylamide concentration. Richards and Temple show that gels which swell are not turbid, and turbid gels do not swell. In practice they confirm that the constant proportion of 5% bisacrylamide (as in Cyanogum 41) leads to satisfactory gels over a wide range of concentration. In agar and agarose, where there are no covalent cross-links, highly clustered gels do not form. At the same time the pore size for 0.5% or even higher concentrations is large, satisfactory for example for electrophoresis of ribosomal RNA (Tsanev et al. 1969). The upper limit of manageability on the other hand is about 5%, so that gels of agar or agarose cannot be used to separate species of low-molecular weight. They are however convenient for high-molecular weight molecules because of their good mechanical properties, Siib,'.' I

ilrclc.,

1'

1s 7

346


and optical characteristics. Most workers use approximately the acrylamide concentrations shown in Table 7.1 for different molecular weight ranges of RNA (the concentration is rarely critical since gels of all concentrations separate nucleic acids over a wide molecular weight range as illustrated for example in Fig. 7.4b). Nevertheless a change of gel concentration can reveal heterogeneity not otherwise obvious. Advantage can be taken of this effect on two-dimensional gels. Fig. 7.13 shows another approach: slices are cut from a 10% gel and subjected to re-electrophoresis at higher concentration and zones are then resolved which are not visible in the first dimension alone. TABLE 7.1 Acrylamide concentration appropriate to different molecular-weight ranges of RNA for satisfactory electrophoretic separations. Molecular weight

Percent acrylamide

Oligomers-10,000

15-20 10 5 2.2-3

10,00G50,000 50,00G200,000 200,000-2,000,000

7.5.5. Choice of buffer Since, as we have indicated, there is in general little to be gained by the use of complex discontinuous buffer systems, almost any buffer functioning between about pH 5 and 9 can be used. Not all buffers give equivalent results however, for there are without doubt interactions between the nucleic acids and the counterions, which in some degree influence the patterns that are obtained. Some scope also exists in the choice of ionic strength. This can be varied within quite wide limits. Low ionic strength (as well as buffer ions of low mobility, such as Tris and similar large organic ions) makes for a large voltage drop across the tubes at a low current; this will minimise heating, and be favourable for rapid separations. Note that the voltage drop across the buffer compartments can also be

Ch. I


347

appreciable under such circumstances, and it should not be assumed that all the applied potential is dropped across the tubes, as would to a first approximation be true in more concentrated buffers). The obverse of these advantages is that the buffering capacity is low, and changes in pH, with migration of ion fronts due to electrode products into the gels, will occur the more readily. Irrespective of concentration, it is desirable to operate at a pH giving maximum buffering capacity for the buffer in question. For reasons that are not entirely apparent, slow runs at high ionic strengths often lead to better separations, a particularly striking example being in the separation of d(AT), oligomers by Elson and Jovin (1969), discussed in 3 9.2.2. It would be generally considered ill-advised to use a pH at either extreme for, in both acid and alkali, slow hydrolysis of the phosphodiester chain must be anticipated, and especially in acid, depurination will also occur. It may be noted however that by entering the pH range of ionisation of bases the possibility arises of creating differences in charge:mass ratio between different species, depending on their composition, which could in principle form the basis of otherwise impossible separations. This strategy has only recently been attempted, and has met with some success, especially when used in conjunction with a conventional pH in two-dimensional electrophoresis (Fig. 7.14b). Composition-dependent change of charge due to partial ionisation of bases supposedly leads also to fractionation in gel isoelectric focussing (Drysdale and Righetti 1972). Of course the true isoelectric point of a nucleic acid would be at very low pH, since to reach it a proportion of the backbone phosphate groups would have to become protonated. The basis of the effects seen in the ‘Ampho1ine’-containingsystems is therefore uncertain, and very possibly involves the formation of complexes between the RNA and the polyampholytes, rather than a result of the pH gradient per se. A further restriction arises when it is required to scan gels directly in the ultraviolet, for then buffers with absorption in this region may not be used. The most common example of such buffers are barbiturate derivatives. It may also be noted that some buffers may liberate Siihjrir side\ p 4x7

348


particularly damaging,noxious and even toxic gaseous products at the electrodes: one such is cacodylate, from which free cacodyl appears to form. Bufferswhich have been widely employed are Tris acetate, pH 7.8 and Tris phosphate, pH 7.7 (Loening 1968). (Tris hydrochloride should be avoided because of hypochlorite formation at the anode.) Buffersmay be made up as concentrated stock solutions and mixed in the correct proportions with acrylamide, agarose, initiator and. catalyst.

7.6. Gel electrophoresis and conformation of nucleic acids 7.6.1. Homology and variability in R N A conformation The use of electrophoretic mobility as a measure of molecular weight is predicated on the assumption that all molecular species involved, both the calibration standards and the unknown, form part of a hydrodynamically homologous series. Such a situation appears to have been substantially achieved for proteins complexed with the detergent SDS (Reynolds and Tanford 1970a): the weight ratio of SDS: protein is a constant for nearly all proteins, and according to the data of Reynolds and Tanford (1970b) the particles of complex can be regarded as ellipsoids in which the minor axis is constant, and the major axis increases smoothly with molecular weight. Thus the retardation by a polyacrylamide gel, that is to say, the mobility, depends only on the molecular weight. A similar result could be achieved in nucleic acids, since the charge:mass ratio is always essentiallyconstant, if conformational differences could be eliminated. That significant frictional differences exist between different RNA species (excluding even the self-evidently different double-stranded molecules that behave more or less as rods) and are reflected in electrophoretic mobility anomalies, is obvious from the results of several workers, in particular Loening (1969) who found that not all RNAs adhered to the same molecular weight calibration, and that their apparent molecular weights interpolated from such calibration changed significantly with ionic strength. Subsequently Groot et al.

Ch. 7


349

(1970) and Grivell et al. (1971) showed that when attempts were made to measure the molecular weight of mitochondrial RNAs by their electrophoretic mobility relative to a calibration set of cytoplasmic ribosomal species, a change in the running temperature from 2" to 28°C led to a 40% increase in the apparent molecular weight. Similar effects were observed in the agarose-acrylamide mixed system by Fisher and Dingman (1971). This must be presumed to be a consequence of a greater temperature effect on the hydrodynamic coil characteristic's of mitochondrial than of cytoplasmic RNA, because of differences in the extent and/or stability of their secondary structure. At the other extreme, attempts to use the gel method to measure the molecular weights of temperature-insensitive unpaired species, such as the poly(rA) tracts at the ends of messenger RNAs, would be grossly in error in the opposite sense if the usual RNA mixtures were used for calibration purposes. Thus, although the electrophoretic mobility in aqueous solvents should always give at least a reasonable rough estimate of molecular weight, and in most cases a fairly accurate value, several workers realised that a more reliable approach would be to eliminate conformational differences by destruction of base-pairing and so reduce all polynucleotide chains to something approaching hydrodynamic homologues. It has of course to be remembered that even a chain devoid of base-pairing is not entirely structureless, since it retains the non-cooperative local stacking interactions between adjoining bases. The work of Eisenberg and FelsenEeld (1967) has clearly shown however that the frictional properties of a single-stranded chain are not greatly affected by stacking until this approaches completion at low temperature, when rod-like behaviour supervenes. Indeed it has been found that fractions of the highly stacked poly(rA) and the totally unstacked poly( rU) fall on the same mobility-molecular weight calibration in aqueous gels at room temperature (Pinder and Gratzer 1974). Thus, within the precision of gel electrophoresis experiments, we are entitled to disregard factors of single-strand stacking, as well as composition-dependent differences in coil flexibility, solvation or counterion fixation. Suhjecr d e , p 4 s 7

350


7.6.2. Electrophoresis in denaturing media An attempt to destroy all base-pairing in RNA by treatment with formaldehyde was made by Boedtker (1971). This appeared to give good results but the apparent promise has not been borne out (Staynov et al. 1972; Boedtker, personal communication) apparently because of partial reversal of the reaction of the formaldehyde with the bases when the reaction mixture is cooled down to the running conditions. Aberrations may also result from the formation of covalent cross-links by this reagent. Reagents such as 8 M urea have also been used to destroy secondary structure, but at neutral pH and room temperature this may not be sufficiently powerful. De Wachter and Fiers (1972) have used 8 M urea in acid solution as a denaturant for RNA in two-dimensional work. This is probably effective, and indeed extremes of pH are in themselves sufficient to break down base-pairs, but in these conditions there are also compositiondependent changes in net charge due to ionisation of the bases. Neutral 8 M urea at elevated temperatures has also been tried (Reijnders et al. 1973) but the procedure is cumbersome and difficult, and produces zones of poor quality. An apparently successful method, which has a number of advantages, and seems to allow unequivocal determination of molecular weights with considerable accuracy was developed by Staynov et al. (1972), and consists in running gels in the non-aqueous solvent, formamide. In this medium both base-pairing and single-stranded stacking are eliminated and electrophoretic separations of high quality are obtained, often, in our hands, better than in aqueous systems. Because of the chemical characteristics of formamide which is an ionising solvent, high activities of hydrogen or hydroxyl ions must be avoided and ionic impurities must first be removed from the solvent by ion-exchange. Formamide thus treated will serve as a medium for polymerisation of the acrylamide, and the gels can therefore be prepared directly in this solvent. Despite the hygroscopic nature of the formamide it does not appear necessary to use formamide rather than water in the reservoirs. In the original formulation of Staynov et al. the supporting electrolyte was sodium

Ch. I


35 I

chloride with no buffer, and external electrodes were used to maintain neutrality and guard against contamination with electrode products. Subsequently the system was redesigned, barbital being used as a buffer. It should be noted that the efficacy of a buffer in a nonaqueous solvent cannot be inferred from its behaviour in water, and that its titration curve must be determined in formamide to establish appropriate conditions for its use. The molecular weight-mobility relations shown in Fig. 7.18 indicate that all RNA species do indeed behave as homologues in formamide. Most strikingly, poly(rA) and poly(rU) fractions conform to the same law. Two-stranded viral RNAs do not necessarily melt spontaneously at room temperature, but remain in the fully melted state after heating to 40 or 50°C (Pinder et al. 1974a). The formamide method has been successfully used for molecular-weight determination of such species (Bevan et al. 1973).

10'

~

0

0.2

04 0.6

0.8

1.0

R f (BPB) Fig. 7.18. Molecular weight-mobility relationships for a series of RNA species ( 0 ) polyriboadenylic 'acid fractions ( 0 )and polyribouridylic acid fractions ( 0 ) in formamide gels (Pinder, Staynov and Gratzer). Siihlm I , d r \ p 487

352


It seems likely that both the accuracy and precision of formamide gel electrophoresis as a method for determination of molecular weights is limited by the uncertainty in the molecular weights of RNA standards (Duesberg and Vogt 1973; Pinder et al. 1974b). With the advent of improvements in ultracentrifuge technique, and particularly a recent method of determination of partial specific volumes (published values of which are startlingly varied), it may be hoped that enough definitive standards will soon become available to cover the entire molecular-weight range, and allow a more exact evaluation of the limits of accuracy of formamide gel electrophoresis. For characteristics of RNA in formamide and a demonstration of its denaturing efficacy, reference should be made to the work of Pinder et al. (1974a). It may be noted that the ultraviolet transparency of good quality formamide allows the scanning of gels directly at 280 nm, but the absorbance rises rapidly below 265-270 nm, and the medium is effectively opaque at 260 nm. 7.6.3. Double-helical nucleic acids Takahashi et al. ( 1969) showed that native high-molecular weight DNA was capable of migrating electrophoretically in agarose gels, and that the mobility was related to the molecular weight in much the same way as for single-stranded RNA species. Given the availability of suitable molecular-weight standards, this seems therefore to offer at once a means of analytical fractionation and molecular weight determination for native DNA. Later Zeiger et al. (1972) showed that DNA would migrate in 2.5% composite polyacrylamideagarose gels, but that in this medium the mobility was more or less independent of molecular weight, though there was a marked dependence on base composition.That molecules in the lo7 molecular weight range should migrate at all in such a medium is in itself remarkable, for they have contour lengths many times the effective pore size, which is in the range of only a few nm (Fawcett and Morris 1966). As to the molecular weight dependence of mobility, it soon transpired that this depended on the size range of the molecules under study.

Ch. I

ESSENTIALS OF GEL ELECTRODES

353

The first systematic study of double-stranded RNA was that of Shatkin et al. (1968) who examined reovirus RNA on 2.5 and 5% polyacrylamide gels, and resolved ten bands, which, based on calibrations by sedimentation analysis, spanned the molecular-weight range 0.8-7.5 x lo6. In both gel concentrations these components evinced a log-linear molecular weight-mobility relationship. Using these same calibrated reovirus RNA components, Wood and Streissle ( 1970) found that eight double-stranded RNA species present in wound tumour virus lay with great precision on the same calibration line. Similar data were obtained for a set of two-stranded viral RNAs by Kawata et al. (1970). The absence of a molecular-weight dependence in the mobility of DNA relates only to molecules of very high molecular weight, and can then be circumvented by irreversible denaturation: Hayward and Smith (1973) working with phage DNAs of 2-30 x lo6 molecular weight per strand found that, whereas no fractionation occurred when the native molecules were run, prior denaturation in alkali gave rise to single strands, which could be well separated and yielded excellent molecular weight calibrations - notably better in fact than obtained for a wide range of single-stranded RNA species. With two-stranded DNA the log (molecular weight) us R, relation is a smooth concave curve (a good example occurs in the work of Danna et al. (1973) on fragments of SV40 DNA), or in the conditions of other workers (Pettersson et al. 1973), a straight line which curves steeply upward in a critical molecular weight range. Up to a molecular weight of about 1-2 x lo6 there is good resolution, and the calibration curve is satisfactory for molecular weight determination, but above this the curvature increases rapidly and resolution vanishes, even though in a gel of sufficiently low concentration the molecules still migrate well enough. Some further unusual features in the electrophoresis of two-stranded species were brought to light by Fisher and Dingman (1971). In the first place, at molecular weights above a critical region, near 4 x lo5, the rate of migration, by contrast with that of the corresponding single-stranded species, became essentially independent of gel concentration over a wide range (at least 1-3% total gel concentration Subject index p. 487

354


Fig. 7.19. Dependence of electrophoretic mobility of native DNA on the applied voltage in polyacrylamide-agarose composite gels, of the indicated acrylamide concentrations. The samples are T2 (@), T1 ( A) and PBSX ( x ) DNA, and fragand T1 ( 0 )DNA (Lishanskaya and Mosevitsky 1973). ments of T2 (0)

Ch. I

355


in mixed polyacrylamide-agarose gels). Secondly, the relative mobilities show a marked dependence on the voltage (Lishanskaya and Mosevitsky 1973). These effects are shown in Fig. 7.19 reflecting the way in which this remarkable effect of gel concentration sets in only at a sharply defined value of the molecular weight. The retardation coefficient, KR merely expresses the slope of the mobilitygel concentration plot, ie. log M = log Mo- KRT, where T is the total gel concentration and log Mo the intercept on the ordinate. The only obvious explanations of these effects is that the native DNA migrates by virtue of orientation along the direction of the potential gradient, and in effect threads its way through the interstices of the gel. The anomalously low retardation in the electrophoretic gels stands in dramatic contrast to the exclusion of similar molecules from molecular sieves in chromatography. The degree of orientation is evidently governed by the magnitude of the voltage gradient, i.e. the orienting force up to some limit, dependent, no doubt, on the rate of rotational relaxation (cf. orientation by flow or electrostatic fields). Moreover at a molecular weight above the critical limit, it is not surprising that the gel concentration has little effect on the mobility, if this really depends only on the ability of the molecule to orient. The dependence of mobility on composition could be related to the degree of flexibility of native DNA (which is generally taken to have a segment length of about 200 A), or on the helix dimensions,which, it has been averred, vary with composition (Bram 1971). According to Dingman et al. (1972b) conditions of voltage gradient and gel concentration can be found (specified in their paper), in which double-stranded species, be they DNA or RNA, in the molecular-weight range 0.02-2 x lo6 fall on the same mobility calibration as single-stranded (i.e. partly paired RNAs), whereas denatured single strands derived from the two-stranded molecules lie on a different curve, most probably because of a low content of the type of internal complementarity that is evidently built into most naturally single-stranded RNAs. The coincidence of the calibrations for the single and double-stranded species has no significance in Subject in&\

0 4x7

356


itself, but is only a chance event for a particular gel concentration, as will be seen. It is not surprising that in staining with a metachromatic dye (in this case ‘Stains-All’) the response of single and double-stranded species is different (Dingman et al. 1970). By densitometric scanning at two wavelengths (corresponding presumably to the aggregated and unaggregated forms of the bound dye), it is possible in principle to distinguish between the two types of conformation. A similar observation, with a visually detectable difference in colour is reported for toluidine blue by Bevan et al. (1973). The conformation dependence of the mobilities of nucleic acids in polyacrylamide gels has been taken a stage further by Dingman et al. (1972a), who have looked also into the effects of cyclic structure. They find that circular single-stranded DNA (which, as has been shown theoretically and experimentally, has a markedly lower frictional coefficient than linear) can be separated from linear DNA of the same molecular weight. Moreover circular duplexes may similarly be resolved from linear duplexes of the same molecular weight. (‘Why circular duplexes migrate at all is not clear in terms of the orientation concept. It is of course possible that the molecule assumes a compact disymmetric form, double helices lying parallel to the long axis, which again is able to penetrate the gel by orientation - a mechanism denied to species of random-coil like character.) Again a single-strand break in a double-helicalsupercoiled circular DNA, which is known to cause relaxation of the supercoil, changes the mobility sufficiently to permit separation from the intact circular form. According to Dingman et al. it is not only the linear duplex DNA, but also single-strandedlinear DNA that (unlike singlestranded RNA) displays a voltage dependence of mobility (Fig. 7.19). Aaij and Borst (1972) have also demonstrated the separation of closed circular DNA double helices from their linear counterparts in polyacrylamide gels, and have found that log-linear calibrations of molecular weight against relative mobility can be set up. A conspectus of the differences in electrophoretic behaviour between three conformational types - linear single-stranded,linear double-stranded and circular double-stranded - is available from the curve relating

Ch. 7


357

mobility and gel concentration obtained by Harley et al. (1973). These profiles depend strikingly on the conformation type and Harley et al. suggest that they could be used to discriminate between them. This claim would seem to be justified if the molecular weight is defined, at least within fairly broad limits. It might be rendered wholly unambiguous by comparing the mobility-gek concentration profile of the native material with that observed after alkaline denaturation (Hayward and Smith 1973), the latter form serving for molecular weight determination, relative to standards in the same state. Harley et al. find that the mobility-gel concentration curves can in all cases be faithfully represented by a second-order polynomial, the coefficients being determined by a standard least-squares curve-fitting routine, i.e. in the form u =a +b T + cT2, where u is the mobility and T the gel concentration. Examples of the characteristics profiles are shown in Figs. 7.20 and 7.21.

-

“

9

:

-

d

7-

VI

5? 6 -

rJ

5 5-

v

.A

c ..z 4-

z

.u

P

3-

g 2I?

4-

8 1. -2 Fig. 7.20. Plot of absolute electrophoretic mobility versus gel concentration for a typical data set, derived here from 23 separate analyses of Mycophage PSl-III, a double-stranded linear RNA. The curve represents the second-order polynomial curve derived from this data set by regression analysis. Subject iiidrx p. 487

358


Fig. 7.21. Dependence of mobility-gel concentration profiles of DNA on conformation. Samples are supercoiled SV40 DNA (sc), nicked circular SV40 DNA (NC) and HeLa cell linear double-stranded DNA. A single-stranded species, 28 S HeLa cell ribosomal RNA is also shown (Harley et al. 1973).

To summarise the salient points: two-stranded nucleic acids migrate in gels and, up to a molecular weight of 1-3x lo6, depending somewhat on the system, can be separated by molecular weight. Above this the mobility depends appreciably on composition but hardly at all on molecular weight, regardless of gel concentration. Single and double-stranded species may be distinguished by measuring mobility as a function of gel concentration, as well as by differences in colour response to the stain. After alkaline denaturation excellent molecular weight-mobility calibrations can be obtained for all molecular weights. Different conformational forms (single and doublestranded, linear, circular and super-coiled) of identical molecular weight can all be separated by gel electrophoresis, and possibly uniquely identified. One may anticipate that some pretty applications of these basic findings will emerge in many branches of nucleic acid research.

CHAPTER 8

The practice of analytical gel electrophoresis of nucleic acids

In this chapter the experimental practice of analytical polyacrylamide gel electrophoresis will be set out. Analytical methods have been evolved to operate on a scale of say 1-10 p g or 1000 counts per minute per component, though this can often be appreciably reduced. We shall not concern ourselves here with true micro-techniques, which function down to the level of the contents of a single cell in some cases, since these are of specialised interest, and have been fully dealt with in a recent and authoritative monograph on micro-methods in molecular biology (Neuhoff 1973). Analytical polyacrylamide gel electrophoresis at the present stage of technology can be made to serve a number of useful purposes. The object may be simply to reveal polydispersity (say degradation) in a sample of RNA. The patterns obtained, however, may be analysed to yield the relative amounts, the kinetics of labelling with radioactive metabolic precursors, and a number of physical properties of the separated species, notably the molecular weights, but touch possibly also on the conformation and effective charge properties of individual nucleic acid species. Gel electrophoresis may also help in the study of molecular structure by controlled degradation or study of the mechanism of action of a nuclease, and of course it is required in the first step of sequencing operations. Selected examples of such applications will be described in ch. 10. We shall aim here to cover such aspects of the technique as are relevant to its utilisation in these various senses. 359

Subject index p . 487

360


8.1. Preparation of R N A Thorough deproteinisation of the RNA is essential for gel electrophoresis; otherwise streaking or even accumulation of material at the origin may be observed. Two procedures described by Loening (1967), and based on earlier work of Kirby and co-workers, appear to be satisfactory. The tissue homogenate containing the RNA to be extracted is first shaken at room temperature with detergent, either 0.5% sodium dodecyl sulphate and 5% sodium 4-aminosalicylate or 2-5% sodium triisopropylnaphthalene-sulphonate.In the first case the suspension is shaken at 0-5°C with phenol containing 0.1% w/v 8-hydroxyquinoline,and, in the second case, the same solution with 10% m-cresol added. The phases are separated by centrifugation at 1000 g for 10 min and the phenol extraction of the aqueous phase is repeated once or twice. The RNA is precipitated at -20°C from the aqueous phase by the addition of 2% w/v sodium acetate and 2.5 vol of ethanol or 3% w/v NaCl and 2 vol ethanol. To remove DNA, the precipitate is dissolved in buffer (0.2 M Tris, pH 7.5)+1.5 M NaCl and a good quality deoxyribonuclease at 10 pg/ml is added. The solution is incubated at 5°C for 30 min, followed by the addition of sodium dodecyl sulfate and deproteinisation with phenol as above. The final RNA is reprecipitated from 0.3 M sodium acetate twice, washed once with ethanol and dried in vucuo before dissolving in electrophoresis buffer.

8.2. Gel components The usual combination of components in the polymerisation mixture, and the only one for which we shall give experimental details comprisesbuffer salts acrylamide monomer, methylene N,N’-bisacrylamide (Bis), tetramethyl ethylenediamine (TEMED) as catalyst and ammonium persulfate as initiator. Variants are possible as mentioned in the previous chapter: other catalysts may be used and also there may in some cases be advantages in eschewing the use of the persulfate, for example, when it is important to avoid too rapid setting of the gel in the course of more than usually complex

Ch. 8

ANALYTICAL GEL ELECTROPHORESIS

361

pouring operations. In such circumstances, photo-initiation by riboflavin is an entirely satisfactory alternative, though, as normally used, it appears to give rise to softer gels, presumably by initiating more, and therefore shorter, chains. Polymerisation ensues only after the solution is irradiated, most simply by a small fluorescent strip-lamp placed a few centimetres away. Details are to be found in Smith (1968). Neither ammonium persulfate nor riboflavin are satisfactory for setting gels in buffer of very low pH, and here Fenton’s reagent can be used (Jordan and Raymond 1969). Another variation of the above procedure is the introduction of hydrolysable cross-linkers in place of the methylene bisacrylamide, when it is desired to dissolve gel slices after the experiment to

8

0

HIC=CH-C-NH-CH,-NH-C-CH=CH,

II

and

0 8 II HIC=CH-C-O-CH2-CH2-O-C-CH-CH2

Fig. 8.1. Hydrolysable cross-linking agents. Exposure of gels prepared with these reagents as cross-linkers to mildly alkaline conditions leads to dissolution. I is diallyltartardiamide(Anker 1970) and I1 is ethylenediacrylate (Choules and Zimm 1965).

release the RNA, e.g. for radioactive counting. Two such have been employed, one being hydrolysable at alkaline pH (too high to allow any hopes that the RNA will not be to a greater or lesser extent covalently damaged) and the other in 2% periodic acid. These are, respectively, ethylenediacrylate (Choules and Zimm 1965) and N,N’diallyltartardiamide (Anker 1970) (Fig. 8.1). They are used in the same way as Bis, but give gels of somewhat different mechanical properties. For the preparation and use of these reagents see the references cited. Both preparations are simple, but ethylene diacrylate is now available from Bordon Chemical Co. Gels from the natural polysaccharide polymers, agar and agarose, can give good results with nucleic acids. They provide satisfactory Subject index p . 487

362

SEPARATION METHODS FOR NLICLEIC ACIDS

molecular sieving in the high-molecular weight range, but are poor media for separating low-molecular weight species, since the highest manageable concentration of gel is not much greater than 4%. We have already noted, and will consider in practical terms, the use of agarose in composite gels to add mechanical rigidity to polyacrylamide of very low concentrations. Agarose is in principle better than agar, since the latter contains sulfate groups, the charge of which leads to a large and undesirable electro-osmotic flow, and the possibility of adsorption effects (Staynov 1972). Not all agarose samples gives good results; satisfactory preparations are those from BDH (British Drug Houses); L'Industrie Biologique FranGaise, and Marine Colloids (called Seakem Agarose). Agarose gels are transparent to ultraviolet light, so the problem of poor UV transmission is eliminated, especially in dried gels in which the scattering characteristic of the gel state, is absent. Beyond the advantages for scanning, Tsanev (personal communication) has described the use of air-dried agarose gels as a matrix for such operations as spectrophotometric titrations. The ionisation state of a species such as RNA in the gel is always that corresponding to the p H of the solution from which it was dried. Thus by immersion in buffers of chosen p H , the agarose film quickly re-equilibrates with the new p H . It is possible by this means to examine, for instance, intact ribosomes in p H ranges in which they would precipitate from free solution. 8.2.1. Purification of gel components Some samples of acrylamide apparently contain large quantities of acrylic acid as an impurity, as well as ultraviolet-absorbing matter, which is apt to vitiate attempts at direct ultraviolet scanning. Some companies(e.g. BDH) now provide a more highly purified acrylamide, which is preferable in this respect. Loening (1967) advocates recrystallisation of both acrylamide and Bis, and this undoubtedly leads to a great improvement in the ultraviolet transparency of the gels. A less demanding, and equally effective procedure is simply to subject the gel to pre-electrophoresis (in the absence of a sample) for a

Ch. 8


363

sufficient time; say 2 hr, with the standard disc gel system. The reservoir buffer is then changed, and samples are applied as usual. It should be borne in mind that the problem of transparency increases with the acrylamide concentration and in low-concentration gels it is often possible to avoid any purification procedure. If desired a substantial improvement in gel transparency can be obtained by soaking the gel in buffer before use. In the case of cylindrical gels this would involve removal from the tube and subsequent replacement - an impossibly tricky operation except for dilute gels. With flat gels there is no problem. We have found pre-electrophoresis to be in general the most satisfactory procedure. Not only do absorbing impurities, which are evidently predominately charged, migrate out of the gel but preelectrophoresis also has the advantage of eliminating the residual ammonium persulfate, which being a strong oxidising agent, is not usually a desirable component of the system. Damage to proteins by ammonium persulfate has been reported, and one could conceive of similar effects when labile minor bases are present, for example, in a tRNA. In discontinuous buffer systems pre-electrophoresis might present problems. The procedure for recrystallisation of monomers follows Loening (1967). 70 g acrylamide are dissolved in 1 1 chloroform at 50°C. The solution is filtered hot without suction. Crystals separate at - 20°C and may be recovered by filtration, washed with chilled chloroform, heptane or both, and dried in air. 10 g of N,N'-methylene bisacrylamide is dissolved in 1 1 of acetone at 4&50"C and filtered hot. The solution is slowly cooled to - 20"C, and the crystals are washed with cold acetone either by centrifugation or filtration. Because both acrylamide and Bis are highly neurotoxic the hot filtrations should be performed in a fume chamber. Care should be taken to avoid skin contact, inhalation of light crystals or ingestion (e.g. by pipetting of monomer solutions). The catalysts T E M E D and DMAEC are likewise very poisonous and additional care must be taken to prevent evaporation into the atmosphere, since they are Subp i index p 487

364


volatile. Once polymerisation is complete it is supposed to be safe to handle the gels.

8.3. Procedures for cylindrical gel electrophoresis 8.3.1. Aqueous polyacrylamide gels 8.3.1.1. Preparation for electrophoresis Preparing the sample Nucleic acids should be in the concentration range 0.01-1 mg/ml and the total loading for analytical purposes is generally of the order of 0.5 pg RNA species. Gels can be overloaded in order to focus attention on minor components sufficiently separated from the bulk material, when the gross spreading of the major zones is of no consequence. In general, however, one should aim to use close to the minimum sample size capable of detection. One of the commonest mistakes of novitiates to the method is to overload the gels. Richards and Lecanidou (1971) have shown that deterioration of resolution occurs in their standard conditions if loads greater than 0.5 pg of a single RNA species are employed in tubes of 0.475 x 7.5 cm. Minimisation of load volume is not, as we have noted ($ 7.5) important: Richards and Lecanidou (1971) have shown that resolution is maintained with a sample column of up to 1 cm in tubes of 6 mm i.d. Into some or all the samples in a run a trace of bromophenol blue is introduced as a marker. Since the best way of applying the sample is by a layer which will remain under the buffer without mixing, sucrose (or glycerol) is added to a concentration of about 5%. Most conveniently 1 tenth of a sample volume of 50% sucrose 0.002% bromophenol blue is added to the sample before application to the gel. Relatively strong zones can also be observed directly during migration by addition of a little stain, such as pyronine Y or acridine orange, but inasmuch as this may cause differential perturbations of mobilities this is probably not advisable in any but fully defined systems. For additional sensitivity acridine orangestained zones can be observed by illumination with light from an ultra-

+

Ch. 8


365

violet lamp (Woods glass filter). Nucleic acids may be protected from the ubiquitous nucleases in a number of ways. The inclusion of small amounts of bentonite (e.g. 100 p g ) is not detrimental if firm gels are used. However, in the absence of agarose, with dilute gels, e.g. 2.4% acrylamide both the gels near the origin and the resulting RNA patterns can be deformed by the bentonite which is particulate and negatively charged. Nucleases can be inhibited also by including sodium dodecyl sulfate (0.1%) in the sample, gel and reservoir buffers, but the detergent must be thoroughly removed by soaking the gel after the run if the usual stains for RNA are to be used, since it interacts with and precipitates the dye. In practice it is difficult to obtain clear and colourless backgrounds, even after extensive soaking of the gel prior to staining. Sodium dodecyl sulfate is transparent in the ultraviolet and therefore does not interfere with ultraviolet scanning of nucleic acids. Hence it is preferable to monitor the distribution of nucleic acids in gels containing SDS by this method rather than by staining with the usual dyes. Detection of a few tenths of a microgram of nucleic acid is possible if it is contained within a single narrow sharp zone in a cylindrical gel of 6 mm diameter cross section, using e.g. methylene or toluidine blue as stains. For adequate signal/noise ratio in ultraviolet densitometry (eg. in the Gilford instrument) a minimum of approximately 0.1 p g RNA in each zone is desirable. Larger amounts allow the use of a lower instrument gain, however, and less noisy traces are obtained. If it is required only to observe the distribution of radioactivity with highly radioactive materials, as is often the case in sequencing operations, ‘invisible’ amounts of labelled RNA may be run. In this case very often carrier RNA is added to the sample at an earlier stage (e.g. before digestion) to assist precipitation; the concentration of carrier must be sufficiently low to keep resolution in the optimal range. The sample may be dissolved in diluted electrophoresis buffer or in water so that, being in a medium of lower conductivity than the gel, it experiences a high field and is rapidly accelerated to form a Sirbjecf index p . 487

366


sharp layer at the gel interface. The bromophenol'blue marker runs in front of all but the smallest oligonucleotides at neutral pH and so serves to indicate the progress of the run. The mobilities of RNA species relative to bromophenol blue are commonly reported. The diffuseness of the bromophenol after it has migrated 2/3 or 3/4 of the length of the gel is an unfortunate aspect of this procedure for comparing samples run in different gels. An RNA marker (e.g. tRNA) provides greater accuracy in determinations of relative mobility (R,). During staining some swelling of the gel may occur. Bromophenol blue distance is measured before staining (in the course of which it diffuses out of the gel), whereas the RNA distance is measured after staining. Thus measurement of the total gel length before and after staining and correction for swelling is necessary for accurate determination of R, relative to this marker. For the precise location of trace amounts of radioactive RNA fragments in gels, De Wachter and Fiers (1971) advocate employing a dye mixture. R, values of the individual dyes in several buffer systems have been published by these authors (Table 8.1). Setting up the tubes Where cylindrical gels are used it is important that tubes should be scrupulously clean. It is recommended that after each run glass tubes should be soaked in chromic acid. They should then be scrubbed with a test-tube brush and detergent, rinsed and carefully dried with ethanol or in an oven. Permanently siliconised tubes (prepared by dipping glass tubes in a 5% solution of dimethyldichlorosilane in chloroform) are said to facilitate cleaning as well as the removal of gels from the tubes immediatelyafterelectrophoresis( Smith 1968).Loening has advocated the use of Perspex tubes, as acrylamide gel adheres less firmly to Perspex than to glass. Adherence to glass is a problem with both very concentrated and very dilute gels which are apt to disintegrate during removal. To set the tubes up for pouring gels, their ends must first be sealed, and they must be mounted vertically. For sealing, two small squares of parafilm may be separately stretched over the bottom and

Ch. 8

367


TABLE 8.1 Electrophoresis conditions and migration distances in vertical slab gels (De Wachter and Fiers 1971). Gel type"

1

2

3

4

Conditions Voltageb,(V) Currentc (mA) Time (hr)

250 20 16

400 30 16

500 25 18

600 15 18

Migration distance (cm) Trypan red Xylene cyanol FF Eosin Bromophenol blue Bromocresol purple Fluorescein Bacteriophage MS2 RNA 23 S ribosomal RNA ( E . coli) 16 S ribosomal RNA ( E . coli) Transfer RNA (yeast) Mononucleotides ~

31.0 -

9.4 10.8 15.5 32.5 Run off

6.0 13.0 18.5 23.0 27.0 32.0

3.2 7.6 10.2 13.5 16.4 20.5

13.8 35

4.5 28

20.0 -

-

~

For the composition of the gels, see Table 8.2 (page 394). Voltage measured between the electrodes. The actual voltage on the gel is about 80% otthis because of the resistance of the wick. Approximate current measured for a 2 mm gel.

twisted around the side of the tube. Plastic caps or rubber bungs may also be used, but the latter introduce a small cavity at the end of the gel, in which an air bubble is easily entrapped. If a ring of plastic tubing is inserted some distance into the bottom of the tube, it may then be closed off with a small glass rod. The removal of the bottom seal requires care, to avoid creating a vacuum which can damage the gel. Tubes are mounted vertically, most simply in a rack in which they are secured, either at the bottom, or near the top by pushing them through a hole in a stiff sheet of rubber or plastic. Subject index p . 487

368


Some useful gel mixtures The monomers are most conveniently made up as concentrated 15-50% stock solutions, containing both acrylamide and bis in the required ratio, in water. They are stable for long periods, often months, though, ultimately, evidently due to factors that have not been investigated, partial spontaneous polymerisation occurs. When ethylene diacrylate is used as a cross-linking agent several authors (Bishop et al. 1967; Peacock and Dingman 1967) advocate the use of up to 3 times higher ratio of cross-linker to acrylamide than normally used in the case of N,N’-methylenebisacrylamide for comparable separations. Concentrated (e.g. 5 x ) buffer stock solutions are made up, and diluted for use both in the gels and reservoirs. The following buffers originally employed by Loening (1967) are commonly used, but numerous others are equally suitable: a) 5 x stock solution of Tris-acetate buffer, pH 7.8 (0.4 M in Tris, 0.05 M acetate, 0.01 M Na,EDTA): Tris base, 24.1 g; Na acetate, anhydrous, 8.2 g; Na,EDTA 2 H,O, 1.85 g; acetic acid to pH 7.8; water to 1 1. b) 5 x stock solution of Tris-phosphate buffer, pH 7.7 (0.36 M Tris, 0.3 M NaH,P04, 0.01 M Na,EDTA): Tris base, 21.8 g; NaH,PO,2 H 2 0 , 23.4 g; Na,EDTA 2 H,O, 1.85 g; water to 1 1. The following solutions were employed by Loening for electrophoresis of RNA in tubes of 2-7.5% acrylamide gels. 5 ml of the stock monomer (15% Cyanogum 41 95 acrylamide: 5 Bis) solution, one-fifth volume 5 x concentrated buffer (see above), and varying amounts of water are mixed to give the desired final acrylamide concentration according to the following schedule: Gel concentration (%) Buffer (ml) Water (ml)

2.0 7.5 24.7

2.2 6.8 22.0

2.4 2.5 6.25 6.0 19.7 18.7

2.6 5.8 17.8

3.0 5.0 14.7

5.0 7.5 3.0 2.0 6.7 2.7

The solution is degassed and then 25 pl TEMED and 0.25 ml freshly dissolved 10% w/v ammonium persulfate are added.

Ch. 8


369

Ribosomal RNA was separated on 2.4% disc gels in 2.5-3 hr at 8 V/cm. At the end of electrophoresis 28s and 18s RNA were found in 2 mm wide bands 2 cm and 4 cm respectively from the origin. The mobilities were somewhat greater when electrophoresis was performed at room temperature, in the presence of 0.1% SDS (gels, sample and reservoir). Pouring the gel In order to ensure that the gels are all of equal length, which is important for comparison of samples within a run, the tubes should have scratch marks for indexing the meniscus. The degassed monomer solution is rapidly delivered with a Pasteur pipette up to the mark, taking care not to splash liquid on the dry part of the tube above the gel solution. A flat interface is then created by overlayering with degassed buffer from a fine Pasteur pipette or as described below. This also serves to protect the top of the gel from atmospheric oxygen which would inhibit polymerisation. Extraction of dissolved air from the monomer solution is desirable for satisfactory polymerisation. Before pouring the gels, the solution may be placed in a small filter flask which is then evacuated on a water suction pump. The solution is allowed to bubble gently for about a minute, and is then ready for use. A very simple procedure, which is particularly convenient when the acrylamide solution is dispensed with a hypodermic syringe, is to draw it into the syringe, close off the end, for example with a pinched-off needle, and then draw back the plunger. This creates a partial vacuum above the liquid, which is sufficient to extract dissolved oxygen. Extra care is necessary in overlayering gels of very low or high acrylamide concentration. If mixing occurs at the new interface, a tacky partially polymerised solution will result at the top of dilute gels. In the case of concentrated acrylamide, mixing to even a small extent will create a solution of appreciable acrylamide concentration in the region of the interface, which will often polymerise in a separate layer. This effect may be avoided by replacing the buffer layer shortly after the concentrated gel is seen to have set, since the dilute solution 5ubjrc.i mdex p . 487

370


will take longer to polymerise. Polymerisation is always attended by formation of a fine refractile boundary almost at the surface of the column. Elson and Jovin (1969) report the use of isobutanol for layering 20% polyacrylamide gels. The isobutanol must be removed soon after setting, otherwise it will dehydrate the gel. Most of these difficulties can be eliminated and near-perfect interfaces between gel and buffer achieved, by the following procedure, recommended by Richards and Lecanidou (1971): the tubes are filled with acrylamide solution to a level about 1 cm above the mark corresponding to the required top of the gel, and overlayered with buffer. A hypodermic needle of the required length is fitted to a syringe and introduced vertically so that the syringe rests on the rim of the tube. The length of the needle should be such that its tip reaches the level required for the top of the gel. The liquid is then carefully sucked off including the original interface. This procedure minimises mixing of the gel and buffer solutions, and creates a very sharp boundary. Gels should set in about 20 min, although longer times must often be allowed for dilute gels, and there is a strong temperature dependence: gels in the cold room will take much longer. Photopolymerisation in riboflavin-containingsolution is induced by placing a fluorescent strip light a few centimetres from the tubes, and is accompanied by bleaching of the yellow riboflavin colour. It is customary to make up gels immediately prior to use, but experience shows that good results can be obtained with gels stored overnight, or even for several days or weeks. Some chemical manufacturers (e.g. Universal Scientific Co., Ltd.) indeed provide ready-made flat polyacrylamide gels which may be stored in contact with their buffer. Immediately before applying the sample, the buffer layer is poured off and the tube above the gel is dried with tissue. The tubes are then inserted into the gel electrophoresis apparatus in symmetrical positions relative to the two electrodes. The tubes are mounted so that the ends project into the two buffer reservoirs. It should be checked that the upper compartment, where the sample is applied, is

Ch. 8


371

the cathode, otherwise disappointment will result. Applying the sample Samples containing sucrose and bromophenol blue are carefully layered on the tops of the gels. The region above the top of the gel should be visible (i.e. above or below the grommets holding the tubes in place), so as to avoid blind layering operations. The tube above the sample is filled with reservoir buffer. Overlayering the sample with reservoir buffer to the top of the tube can be done with a Pasteur pipette with the tip drawn out to a fine point (preferably slightly bent) or with a syringe with a needle bent at right angles (in bending a syringe needle, a thick wire is first inserted into the hole: otherwise the needle wili be blocked at the bend). The buffer is gently expelled against the wall of the tube. The presence of bromophenol blue in the sample helps to detect any mixing, or one may watch that the refractive index boundary between the sample solution and buffer remains sharp. Alternatively the sample may be layered under the reservoir buffer. The sample volume is not critical, as mentioned above, but should be no larger than necessary. For gels of 0.6 cm diameter, 0.054.15 ml is a convenient volume, but up to 0.5 ml may be used without great deterioration of the results if the ionic strength is substantially lower in the sample than in the reservoir and gel buffers. After filling the tubes with buffer, the upper reservoir may be filled (e.g. through a funnel with its end directed away from the tops of the tubes) without fear of disturbing the sample. If air bubbles are lodged under the gels in the lower reservoir they should be dislodged with a bent Pasteur pipette. 8.3.1.2. Electrophoresis In discontinuous buffer systems especially (Davis 1964; Ornstein 1964), the conductivity of the gel varies through the run. For this reason constant-current power suppliesare always advocated. Though the resistance always varies somewhat, probably largely while the gel warms up to its equilibrium running temperature, it is not our Subject index p . 487

372


experience that the use of constant voltage makes any appreciable difference to the appearance of the electrophoretic patterns. In any case, the voltage should be adjusted after perhaps 15 min, and thereafter, at least until serious chemical changes due to electrolysis occur, it will remain reasonably constant (typically at 5 mA per tube, for 8, 6 mm diameter tubes in the Tris-acetate buffer described above). 8.3.1.3. .4. fter elecrroplioresis Removing the gel As the concentration of acrylamide increases, gels become increasingly difficult to remove from tubes. In general it is necessary to loosen the gel from the sides of the tube, and this can be accomplished by means of a long thin hypodermic needle or a waterfilled syringe, taking care not to scratch the gel. The tube is rimmed with a needle, at both ends if necessary, while water is passed through the needle. Some pressure, depending on the gel concentration, is then necessary to expel the gel from the tube. In the case of dilute gels in Perspex it may be sufficient to remove the retainer (plastic ring or gauze), which it is prudent to attach to the bottom of tubes made from this material, and allow the gel to slide out. A rubber pipetting teat filled with water may be used to encourage the process.For stronger gels, the tube can be connected via rubber tubing to a water tap and the gel forced out by gentle water pressure. This requires practice. Thejunction of the rubber tubing and the gel tube is held firmly with one hand, and the water pressure is controlled with the other, being turned down when the gel begins to move. Application of a rubber bulb full of water is a good alternative. The gel should not be ejected too rapidly. It tends to break especially at the point passing the end of the tube (it is important that the end of the tube be smooth) and with insufficient care, the gel may emerge explosively and disintegrate. It is advisable to eject it into a beaker of water. For very stiff gels, Johns (1967) advocates cracking the tube in a vice, and peeling away the glass. Published photographs of gels pieced together from the fragments tell their own tale. It is worth mentioning that gels of very high acrylamide concentration

Ch. 8

313


become more manageable if say 15% glycerol is included in the gel buffer as a plasticiser.

Staining and clearing Dyes in common use for staining nucleic acids in polyacrylamide gels are acridine orange, methylene blue, toluidine blue and pyronine Y , although perusal of the literature on histochemical stains (Conn 1946; Gurr 1962) suggests that many others might be equally suitable. Even very small oligonucleotides (containing 10 nucleotide residues or even fewer) can be stained with acridine orange (Gould et al. 1969). The actual limiting factor may be the precipitability of the RNA rather than its ability to adsorb the various dyes. This may be expected to set in around 5 nucleotides in the conditions used (e.g. 15% acetic acid). Dahlberg et al. (1969) made use of a dye which they termed ‘Stains-all’ ( 1-ethyl-2-[3 -( 1-ethylnaphtho [1,2d]thiazolin-2-ylidene)-2methylpropenyl]naphtho[ 1,2d]thiazolium bromide from Eastman Organic Chemicals) because it stains RNA (bluish purple), D N A (blue) and protein (red). In the procedure as used in the authors’ laboratories the gels are immersed in a 2% solution of pyronine Y (or acridine orange) in 15% acetic acid for several hours; cylindrical gels are in test tubes, flat gels in developer trays. Background dye is removed by washing the gel in 5-10% acetic acid. Other workers using methylene blue and toluidine blue (Peacock and Dingman 1967)prefer the procedure of immersing the gel first in a solution of 1 N acetic acid for 15 min to precipitate the RNA, and then in a solution of 0.2% dye in 0.04 M acetate buffer at pH 4.7 (equal parts of 0.4 M sodium acetate and 0.4 M acetic acid) for 1 hr. Background dye is removed by several changes of water. It is best to stop short of complete destaining when there is still a trace of free dye in the bathing medium in equilibrium with the RNA, since otherwise the stain may also be stripped from the RNA. The RNA itself does not appear to redissolve and it is usually possible to restain it with fresh dye. Several types of apparatus for electrophoretic destaining have been devised, and two designs have been shown in Figs. 7.15 and 7.16. In Sublet t index

p. 487

374


the first (Richards and Gratzer, 1968) (obtainable from Shandon Scientific Co.), the gels rest in a compartment bounded by porous plastic, e.g. 1/32" thick Vyon (obtained from Porous Plastics, Ltd.). The plastic separates the gels from the electrode compartments containing carbon rod electrodes and 5-10% acetic acid. The rods are connected across a 50 V D.C. supply which passes about 0.5 A through the membranes and the gels, It is recommended to destain pyronine Y in a cold room. The second apparatus (Fig. 7.16) is rather simpler, since apart from depositing the gels in the slots provided there are no further assembly operations. In electrophoretic destaining dye may be removed from RNA zones; too high a voltage must be avoided, and patience exercised. In fact, many workers prefer to avoid the risk of stripping dye from zones, and simply change the destaining fluid at intervals until the background is sufficiently clear. The process can be accelerated by placing the gels into slots in a Perspex frame in a dish in which they can be covered with liquid and.mounting this on a rocking table, so that a gentle agitation can be maintained, alternatively a 'fish tank' water circulator can be used.

8.3.1.4. Evaluation of non-radioactive components Stained gels can be kept for long periods in dilute acetic acid. Storage in the cold room with a drop of chloroform added to the medium discourages the growth of mould. The gels are best examined against a background of diffuse light. For this purpose a light box (e.g. Kodak 'Cold Light' illuminator) is appropriate. Photography Gels are mounted on a uniform source of diffuse white light, e.g. a Kodak 'Cold Light' illuminator. Flat gels are placed in a clean Perspex dish over the light and tubes containing disc gels are mounted vertically on a frame against the light box which stands upright. The gels must be immersed in storage buffer. The camera is placed at a distance permitting the gels to be brought in focus. A viewing camera has the advantages that the gels can be seen clearly, and good development of the film is facilitated by virtue of the large

Ch. 8


315

size of the negative. Kodak Panchromatic film is recommended for gels stained with blue dyes (methylene or toluidine blue) in conjunction with a Kodak Wratten No. 58 filter (plus a No. 22 filter for increased contrast). The films are developed with Kodak DGlO Developer (diluted 1:4) for 4 min, or for lower contrast (when zones of widely varying intensities are to be recorded), May and Baker ‘Qualitol’ (diluted 1:9) for 4-5 min. Similar conditions are used for direct photography of unstqined RNA in gels illuminated with UV light in a dark room. Polaroid film is also adequate for the photography of blue-stained gels and for UV photography, but not for gels stained with acridine orange or pyronine Y (magenta). In the latter case Ilford Orthoset film is recommended. If the zones are faint a No. 58 Wratten filter can be used. Otherwise no filter is needed. DG-10 Developer is used as described above. A rational approach to the choice of filters is to select one, for example from the Kodak Wratten Filter manual, with a window showing maximal transmission at the absorption maximum of the nucleic acid-dye complex. However the results obtained are probably not primarily dependent on the filter and, if a very exacting photographic record is required, extensive experimentation with film and development and exposure times is generally unavoidable. Densitometry In preparing the sample for densitometry, it is essential to ensure that it lies at the beam focus, and in the case of cylindrical gels it is generally considered advisable that a flat surface is presented to the beam to prevent error from lens effects in the gel. This is best accomplished, as in the Gilford arrangement (Fig. 7.17) by slight compression of the cylinders between the walls of a silica trough. If the zones are distorted - for example by the ‘funneling’ effect that occurs when there is excessive heating during electrophoresis - errors may be expected. Moreover, surface imperfections, if the gel has been damaged during removal from the tube, or bubbles, can lead to artefacts. If there is any question of whether excursions in the densitometer trace are genuine, it is prudent to repeat the densjtometry at a long wavelength, outside the absorption band of the Subjecr index p . 487

316


dye, at which only the surface artefacts will be recorded. Adherence to Beer’s Law over considerable ranges of RNA concentration has been reported by various authors. I t should benoted that stacking dyes, such as are usually used in staining, do not of themselves obey Beer’s Law. It is reasonable to expect that at saturating concentrations of dye on a nucleic acid, Beer’s Law will be obeyed, but there may be some degree of dependence on the composition of the nucleic acid. It is probably advisable for anyone concerned with precise measurements of relative concentrations in gels to check that under their experimental conditions, and with their materials, Beer’s Law is in fact obeyed. Finally, it should be ensured that the scan rate is such that the recorder can follow the changes in absorbance without lag. Relative concentrations are obtained by integration of the areas under the peaks, e.g. by planimetry, or less exactly by cutting out and weighing.

Dye-elution The estimation of eluted dye constitutes an alternative method of quantitation. Zones, and samples of background (ie. gel segmentscontaining no zones) are cut out with a razor blade or scalpel. The pieces are macerated, for example in a Potter homogeniser, with a little salt solution, saturated with urea. The debris are separated by centrifugation, and are re-extracted. The extracts are bulked and either weighed or made up to a standard volume, and their absorbance is measured in a spectrophotometer. A good procedure (if RNA is not to be recovered) depending on the chemical stability of the dye, is to extract with 0.5 N alkali, when most dyes of the acridine type become detached from the nucleic acids, and elute readily from the gel matrix. The absorption maximum should be located in the spectrophotometer, and used for comparative concentration determination. At neutrality or in the absence of urea, it may be advisable to guard against deviation from Beer’s Law by selecting a wavelength corresponding to the isosbestic point for the self-association of the dye (e.g. 492 nm for acridine orange).

Ch. 8


377

8.3.1.5. Evaluation of radioactive components As we have noted (5 7.5.3.4), the two approaches are autoradiography and slicing, with or without extraction, followed by counting. For RNA the most usual isotopes are 32P, 14C and 3H; 35S can be incorporated into the minor thiolated bases, present in tRNA. Autorudiogruphy For autoradiography of ’P-labelled RNA, wet gel slices may be used, since the high energy of the /3-particles emitted enables them to penetrate the gel. Autoradiography of 14C on the other hand requires a thin flat slice of dry gel. The following procedure is described by Fairbanks et al. (1965): the gel is placed in a specially constructed holder (Fig. 8.2) and sliced longitudinally into 4 slices by 3 taut wires in a frame, the two inner slices (1/16” thick with flat faces) being suitable for autoradiography. The slice is

Fig. 8.2. Device for preparing flat longitudinal slices of cylindrical gels for autoradiography. The block (lower left) is made out of Perspex. The clearance between the walls and the spacers, and between the spacers themselves (expanded drawing at right) is sufficient to allow the framed taut steel wires (upper left) to be drawn through. This provides two flat sections of gel (Fairbanks et al. 1965). Subjecr index p . 487

378


placed on Whatman 3 MM filter paper, which is then laid on a rigid piece of porous plastic, e.g. Vyon, fitting over the inverted top part of a small vacuum desiccator with the suction outlet at the bottom. While still- wet the plastic with the gel slice(s) is covered with polyethylene sheet or Saran wrap, which is secured with a rubber band around the desiccator to make an air-tight seal. The water is removed from the gel by gently sucking air out through the bottom of the desiccator. An infrared lamp placed over the gel may help the process, but too rapid dehydration can lead to cracking of the polyacrylamide film. This procedure is applicable only to gels within a limited range of concentrations. Very dilute gels (3 or 4%) cannot be sectioned properly because they are not sufficiently rigid. The addition of agarose however overcomes this difficulty. Gels of 10% acrylamide or higher are brittle. Soaking the gel in glycerol solution as plasticiser has been suggested (Gordon 1969). The use of lower cross-linker concentrations, so that the gel swells on soaking may also help. A variant of the technique described by Adesnik (1973) entailing the use of an atmosphere of steam over the gel slice permits the drying of concentrated acrylamide without cracking. Fairbanks et al. suggest that drying should require a few hours. Radioactivity levels of 1000 dpm I4C in a zone of 0.25 xO.6 mm required a day’s exposure to X-ray film to obtain a blackening sufficient for quantitation by densitometry. Slicing the gel Three types ofdevice have been used to obtain uniform horizontal sections of cylindrical polyacrylamide gels which can be used for measuring radioactivity. A frame containing wires at fixed intervals (e.g. 1 mm) can be used to cut the gel horizontally (Fig. 8.3). Secondly the gel may be sectioned with a microtome device, cutting at fixed intervals. Thirdly the gel may be extruded through a small orifice or wire mesh and fractions eluted into a fraction collector (or collected by hand). A microtome is specially manufactured by Mickle Co., an adaptation of their tissue-slicing microtome, for slicing cylindrical polyacrylamide gels. It differs from the original microtome in having

Ch. 8


379

I

Fig. 8.3. Device for slicing cylindrical gels. The gel lies in the cylindrical groove in the brass block with grooves sawn at 1-mm intervals. The frame (above) has steel wire drawn round studs at both ends to give a separation equal to that of the grooves in the block. The wire grid so formed is drawn through the gel to give mm-thick discs (Chrambach 1966).

Fig. 8.4. Gel fractionator (Maize1 1966). Subiecr index p. 487

380


a stainless steel tray surrounded by troughs with an elevated platform in the middle for the gel. The troughs are filled with ground dry ice which hardens the gel, to allow cutting. Thus, very dilute and easily deformable gels may be sectioned with ease. The gel should be fixed between rubber stops since freezing is apt to cause expansion. This method of slicing is also applicable to very concentrated gels (a15% acrylamide) without freezing, which is useful since other methods (e.g. Maizel's fractionator described below) are not suitable for such gels. Indian ink markers can be injected into the gel as an aid to relating migration distance and radioactivity. The third type of device (Fig. 8.4) originated by Maizel (1966) in order to section cylindrical polyacrylamide gels, employs two motordriven pistons which simultaneously force free buffer and the gel, contained in a stainless steel cartridge, through a narrow orifice (0.125 mm), or in the commercial version a wire mesh with the dimensions of an electron microscope grid, into a mixing chamber and out to a fraction collector. This device is specially convenient for radioactive samples, since the fractions can be collected directly into scintillation vials. A specially designed fraction collector to hold scintillation vials is manufactured to accompany the gel fractionator (Savant Instruments, Inc.). In an experiment with radioactive protein uniformly distributed through a 10 cm gel, Maizel reports collecting 60 fractions in which the radioactivity shows a variation of not more than 10%. Radioactive counting Fractions containing 32Por I4C label may be counted either in a low-background gas-flow counter or in a scintillation counter. Only the latter method is sufficiently sensitive to deal with the weak p-emission of 3H compounds. Slices of gel can be dried on adhesive labels and transferred to planchets for counting in a gas-flow counter. Loening (1968) describes a method for processing a large number of samples. Slices are dried on 16 mm cine film, placing one slice on every other frame. The film is fed under a Geiger tube using a cine camera or projector mechanism and the counts for each slice are thus printed out.

Ch. 8

38 1


Loening ( 1968) advocates counting 'H-labelled RNA by putting slices of gel in 0.5 ml 6% piperidine base in open vials, and allowing them to dry in air. 0.5 ml water is added to each causing the gel to swell, followed by 12 ml of a water-soluble scintillation fluid. Some 90% of the RNA is extracted. Young and Fulhorst (1965) make use of the fact that polyacrylamide gels will dissolve in H,O,. Terman (1970) puts 1 mm slices of gel into tightly capped nylon scintillation vials with 0.3 ml NCS TM solubiliser (Amersham/Searle Corp.). The vials are shaken for 10-20 hr at 3345°C. Swollen slices (about 4-fold increase in volume) could then be fully solubilised in toluene. 10 ml toluene containing 6 g/1 Permablend (Packard) is added to each vial with swirling. Counting, performed with either a Nuclear Chicago Model 772 or Beckman 15-200 B Liquid Scintillation Spectrometer was achieved with an efficiency of 49% for 3H. Zaitlin ( 1970) reported low counting efficiency for concentrated gels by Terman's procedure and accordingly modified it to obtain higher efficiency. l-mm slices of gel are placed in the bottom of glass scintillation vials; then 0.7 ml of a solution containing NCS diluted 1:9 v/v with water is added. Vials are tightly capped and heated overnight in an oven at 50T, causing the gels to swell. The vials are tilted slightly during heating to ensure contact between the gel slice and liquid. To each vial is added 10 ml scintillation mixture containing 0.5% 2,5-diphenyloxazole and 0.01% 1,4-bis-2(5-phenyloxaxo1yl)benzene (Packard Instrument Co.) in toluene. The vial is swirled to mix the contents and is then ready for counting. Gels containing hydrolysable cross-links in place of N,N'-methylenebisacrylamide offer advantages for counting. To dissolve ethylene diacrylate cross-linked gels (Choules and Zimm 1965), l-mm slices are placed in the bottom of 20 ml scintillation vials, and the digestion is started by adding 0.5 ml of a solution containing 9 vol of 1 M piperidine and 1 vol alcoholic Hyamine solution. The gel dissolves in 1-4 hr, depending on th'e acrylamide and diacrylate concentrations. The process is hastened by placing the vials on a shaker at 37°C. Following digestion, 10 ml Kinard's .scintillation fluid is added and mixed to give a clear homogeneous solution. The piperidine conS"b,@

Laboratory Techniques in Biochemistry and Molecular Biology Vol 04

Laboratory Techniques in Biochemistry and Molecular Biology Vol 04

Laboratory Techniques in Biochemistry and Molecular Biology Vol 12: Photogenerated Reagents in Biochemistry and Molecular Biology

Laboratory Techniques in Biochemistry and Molecular Biology Vol 05

Laboratory Techniques in Biochemistry and Molecular Biology: 2 Parrts

Laboratory Techniques in Biochemistry and Molecular Biology Vol 33: FRET and FLIM Techniques

Analytical Techniques in Biochemistry and Molecular Biology

Laboratory Techniques in Biochemistry and Molecular Biology Vol 27: Ultrasensitive and Rapid Enzyme Immunoassay

Laboratory Techniques in Biochemistry and Molecular Biology Vol 26: Analysis of RNA-Protein Complexes 'in vitro'

Laboratory Techniques in Biochemistry and Molecular Biology Vol 21: Isolated Hepatocytes Preparation, Properties and Applications

Laboratory Techniques in Biochemistry and Molecular Biology Vol 15: Practice and Theory of Enzyme Immunoassays

Laboratory Techniques in Biochemistry and Molecular Biology Vol 07: Preparation and Mammalian Plasma Membranes

Laboratory Techniques in Biochemistry and Molecular Biology Vol 28: Synthetic Peptides as Antigens

Laboratory Techniques in Biochemistry and Molecular Biology Vol 31: Adeno-associated Virus Vectors for Gene Therapy

Laboratory Techniques in Biochemistry and Molecular Biology Vol 29: Cancer Metastasis: Experimental Approaches

Laboratory Techniques in Biochemistry and Molecular Biology Vol 10: DNA Sequencing

Laboratory Techniques in Biochemistry and Molecular Biology Vol 08: Cell Culture for Biochemists

Molecular Biology Techniques: An Intensive Laboratory Course

Molecular Biology Techniques: An Intensive Laboratory Course

Techniques in Molecular Biology

Insect Molecular Biology and Biochemistry

Avian Biochemistry and Molecular Biology

Laboratory Techniques in Biochemistry and Molecular Biology: Isoelectric Focusing - Therory, Methodology and Applications

Laboratory Techniques in Biochemistry and Molecular Biology: Guidebook to Lipoprotein Technique

Principles and Techniques of Biochemistry and Molecular Biology

Human Molecular Biology Laboratory Manual

Human Molecular Biology Laboratory Manual

Human Molecular Biology Laboratory Manual

Techniques in Molecular and cellular Biology

Oral Biology: Molecular Techniques and Applications (Methods in Molecular Biology, Vol. 666)

Laboratory Techniques in Biochemistry and Molecular Biology Vol 04